Articles

• 18.11.2025
• ~121 min.
• Juan García

The Technological Mirage of the Twenty-First Century

The Great Artificial Intelligence Bubble
Anatomy of a Foretold Bubble

1. Introduction

In recent months I have watched, without being able to do anything about it, how hype, sensationalism, and above all human ignorance have buried the potential of a technology that, although destined to shape our near future, may destroy it even before it arrives.

The hypocrisy of many self-proclaimed “AI experts” is leading the vast majority of people to get lost among lies and false promises in a field called “generative AI”, wrapping them in terms like “AI Agent”, “agentic AI”, using titles such as “I have created the best AI agent that does EVERYTHING for you with just one agent” or “this AI agent will make you millions” or “I am earning thousands of euros with an AI agent”; calling for action with comments like “if you want this AI agent, comment AGENT and I’ll send it to you by PM”.

Hypocrisy and ignorance, two deadly weapons capable of destroying even the strongest society in the deepest silence.

As you may already know, OpenAI launched the new 5.1 model this week. I read Sam Altman (CEO of OpenAI) on X (formerly Twitter) saying that this model was intelligent. Honestly, my stomach turned when I read this.

Although it does not surprise me, I still cannot believe the situation we are living in. For all these reasons I have decided to take my time to show you that the AI bubble does indeed exist, and that if we do not take the necessary measures and choose the correct path as soon as possible with respect to this technology, this bubble will explode in our faces, and its consequences will be irreversible for decades. Writing this article has taken me a great deal of time and effort, and I sincerely hope you value it.

We are heading toward a crisis unlike any other in history, and no other era has invested so much money, nor had so many platforms and hypocrites spreading hype driven by so much sensationalism in the way it is happening today.

In recent decades the term “artificial intelligence” (AI) has gone from being a technical promise to a central element in the public, industrial, and financial agenda. Today we face the widespread conviction that AI will radically change the global economy, business models, the labor market, and social structure. However, this conviction is not free from risk. In this article I argue that beyond a technological revolution, AI is immersed in what can be defined as a bubble: a speculative phenomenon driven by excessive expectations, massive investments, stock market valuations that no longer reflect tangible results, and a dominant narrative presenting AI as the definitive solution to all human challenges. Considering my experience as an author and trainer in AI, this issue requires a critical and nuanced view that identifies both the real potential and the shortcomings, the excesses, and the vulnerabilities of the current environment.

Why claiming that we are already inside this bubble deserves attention. According to recent data, for example, a survey of fund managers by Bank of America Global Research shows that approximately 54 percent of investors believe that AI-related stocks are already in a bubble. Likewise, an analysis by Deutsche Bank indicates that although OpenAI’s revenues in the first half of 2025 were around 4.3 billion dollars, its net loss was estimated at 13.5 billion, more than triple in losses. Additionally, a study by MIT (Massachusetts Institute of Technology) reports that 95 percent of generative AI initiatives did not generate a return on investment. These figures point to a growing disconnect between investments, expectations, and real results.

To distinguish this phenomenon from a mere technological fad, it is necessary to have a rigorous definition. A technological fad refers to a period of elevated and temporary interest in an emerging innovation, which may generate rapid adoption but also setbacks without necessarily causing a systemic crisis. A bubble, on the other hand, implies not only enthusiasm and accelerated adoption, but also the existence of inflated valuations that are not justified by economic fundamentals, massive investments based on future expectations rather than present utility, and an underlying risk of abrupt correction that may lead to a collapse of value, widespread losses, and chain effects on the economy as a whole. In the case of AI we are seeing a combination of these three factors: accelerated hype, valuations that seem disproportionate, and a volume of capital and infrastructure growing at an unsustainable pace.

Every day I have to deal with problems that self-proclaimed “AI experts” have caused and are causing to companies. And I say every day, not every week or every month. Every day. And it is getting worse. This happens in Spain, Germany, Switzerland, the Netherlands, Austria, France, the United Kingdom, and countries in the Americas. Problems ranging from 9,800.00 euros to 246,798.59 euros. All due to poor advice and total ignorance from these “AI experts”.

The relationship between expectations, promises, and technical reality stands out as the core of the problem. Promises that AI will enable universal automation, superintelligence, the replacement of human labor across a wide range of functions, or the creation of entirely new massive industries have become dominant. But the technical reality does not yet support many of these claims. For example, recent empirical results show that developers using AI tools completed their tasks 20 percent slower than without them. Likewise, the infrastructure costs associated with AI (data centers, high-performance chips, energy) are already estimated in the hundreds of billions of dollars, while the current market for AI services remains much more modest. This creates a growing gap between what is promised (and imagined) and what is delivered in terms of return, efficiency, scale, and real transformation.

The disconnect between real economic value and stock market or media valuation is another clear sign that we are in a bubble. Companies labeled “AI-first” or “AI-native” receive large investments and astronomical valuations, in some cases without a proven business model or significant revenue. For example, some firms have been valued in the tens of billions before demonstrating profitability. At the same time, the media and public narrative feed grand expectations, which in turn fuels speculation and reduces the weight of technical or realistic analysis. When the market anticipates that all future value streams are already priced into the present, any sign of disappointment or delay can trigger a severe correction.

The paradox of the “boom of the century” lies in the fact that AI is indeed a technology with transformative potential for productivity, services, medicine, industry, and society as a whole. But at the same time, this boom risks imploding due to an excess of faith. When the dominant narrative replaces critical analysis and massive investment is directed toward expectations rather than realities, a fragile ecosystem is created that may collapse if revenue lines or returns do not appear. The conversation about AI no longer concerns only what is technically possible, but how we manage euphoria, economic narrative, regulation, education, and institutional oversight. As an AI expert and trainer, I am in a privileged position to warn about this risk and to bring clarity to a transformation that cannot rely solely on faith, but requires technical robustness, rigorous business models, and a critical mindset.

1.1. What we mean by the artificial intelligence bubble and why stating that we are already inside one is justified

The concept of the artificial intelligence bubble must be understood as an economic, technological, and sociopsychological phenomenon that combines inflated expectations, accelerated investment, and a disproportionate valuation compared to the real value created. In financial terms, a bubble occurs when the price of an asset or an entire sector drifts significantly away from its economic fundamentals due to speculation and collective investor behavior. In the case of AI, this phenomenon is amplified by the intangibly seductive nature of the technology, which is perceived as a force capable of redefining civilization and altering the global economic order.

Stating that we are already inside an AI bubble is not an ideological declaration, but an empirical reading of multiple indicators that align with the classic overvaluation patterns observed in previous bubbles. The first of these is the disconnection between tangible results and financial projections. According to data published by Bloomberg and Deutsche Bank, global spending on infrastructure and AI model development will exceed 200 billion dollars annually in 2025, an increase of more than 800 percent in just three years. However, fewer than 15 percent of the companies that have implemented AI solutions report a positive or measurable economic return. This gap between investment and profitability is a characteristic symptom of growth driven by belief rather than evidence.

Another key indicator is the concentration of value in a very small number of companies. Nvidia, for example, currently represents more than 30 percent of the total value of the US technology sector linked to AI, with a market capitalization surpassing 2.8 trillion dollars in October 2025. This means that a single company carries a large part of the global growth narrative, a situation comparable to the role Cisco Systems played during the dot-com bubble in the year 2000. At the same time, hundreds of startups that define themselves as AI companies generate no profits, depend on venture capital, and operate with business models that do not guarantee sustainability.

In parallel, financial analysts have begun to warn of a narrative concentration risk. The International Monetary Fund published a report in September 2025 acknowledging that the enthusiasm for artificial intelligence is altering market structure, diverting capital away from productive sectors toward purely speculative projects or those dependent on long-term growth expectations. This means that AI is not only reshaping economic discourse but also the global allocation of resources, increasing vulnerability to a potential market correction.

Another element that shows we are inside a bubble is the media and conceptual inflation of the term AI. Any application using classification algorithms, recommendation systems, or simple statistical automations is promoted as artificial intelligence. This is one of the biggest problems, as this phenomenon generates a false perception of technological omnipresence, creating the illusion of broader progress than actually exists. Technically, a large portion of current implementations are language or vision models trained through supervised learning, without understanding or autonomous reasoning, and highly dependent on preexisting data. The gap between what the public believes AI is and what it truly is widens each day, feeding a climate of irrational expectations.

Additionally, the operational costs associated with the development and maintenance of AI are beginning to show their structural weight. Data centers dedicated to training generative models consume millions of liters of water per year and massive amounts of electricity. According to the International Energy Agency, global energy consumption of AI systems could reach one thousand terawatt hours in 2026, equivalent to the consumption of the entire United Kingdom. Despite this, a large part of current investments does not account for environmental sustainability or long-term profitability, which increases systemic fragility.

The perception of inevitability and technological salvation, characteristic of major historical bubbles, dominates today’s discourse. The belief that AI will solve all problems—from productivity to climate change—has generated a messianic narrative that replaces critical analysis with blind faith in innovation. Large consulting firms, media outlets, and social networks amplify this message, encouraging governments, institutions, and companies to invest without rigorously evaluating the impact or viability of the projects.

For all these reasons, stating that we are already inside the AI bubble is not an exaggeration, but the recognition of a recurring historical pattern: a cycle of technological euphoria, inflated expectations, misalignment between investment and return, and a gradual loss of critical sense. As with all previous bubbles, the question is not whether it will burst, but when and with what consequences for the economy, politics, and global society.

1.2. How it differs from a simple technological fad

A technological fad can be defined as a short-lived trend that emerges when an innovation sparks a level of social or media enthusiasm disproportionate to its maturity or real impact. These kinds of phenomena are recurrent throughout the history of technology and are usually characterized by rapid initial adoption, excessive expectations, and subsequent stabilization or decline when the results fail to meet the original promises. Recent examples include cryptocurrencies, the metaverse, or augmented reality, all of them technologies with solid technical foundations but whose public projection was distorted by hyperbolic marketing and an exaggerated perception of immediate transformative power.

The artificial intelligence bubble, however, presents structural differences that place it on another level. It is not a passing trend but a systemic phenomenon that simultaneously involves the financial sector, governments, the tech industry, and society as a whole. The magnitude of investment, the concentration of economic power, the scale of institutional adoption, and the dependence that has already formed around AI far exceed the parameters of a simple trend. According to data from the International Monetary Fund and the consulting firm McKinsey, global investments in AI infrastructure and development will surpass 380 billion dollars in 2025, an amount equivalent to almost twice the annual GDP of Finland. This volume of capital does not respond to a temporary curiosity from the market but to a structural commitment that already ties up public and private resources on a global scale.

Another key difference lies in the nature of the expectations generated. In a technological fad, the enthusiasm focuses on product adoption or on fascination with novelty. In a bubble, expectations shift toward the future: investment is made not in what the technology does today, but in what it is believed it will be able to do tomorrow. This temporal displacement of value is what multiplies the risk. In today’s AI landscape, most market valuations are not based on performance metrics or cash flows but on promises of economic transformation that have not yet materialized. Companies that barely generate revenue are valued in the billions of dollars because they are perceived as carriers of the future.

In addition, a technological fad may remain confined to a consumer segment or niche, whereas the AI bubble cuts across all productive and administrative sectors. AI has become a symbolic requirement for modernity and competitiveness. Banks, hospitals, governments, universities, and media outlets incorporate artificial intelligence projects even without a clear technical or economic justification, simply for the sake of not falling behind. This institutional contagion effect, a form of strategic imitation, is characteristic of bubbles rather than fads.

There are also differences in the degree of macroeconomic impact. A technological fad may fade without significantly altering the stability of global markets. A bubble, on the other hand, has the potential to destabilize entire sectors. AI’s dependence on the semiconductor supply chain, the massive investment in energy infrastructure, and the hoarding of computational resources are generating geopolitical and economic pressures that are already affecting global balance. According to The Economist, global spending on high-performance chips grew by 440 percent between 2022 and 2025, driven almost exclusively by demand from the artificial intelligence sector. This increase has pushed prices up, created bottlenecks in production, and caused distortions in other industries that depend on the same hardware.

Technological fads tend to fade gradually as interest wanes or the public becomes saturated. Bubbles, by contrast, do not simply dissolve: they burst. Their end is not a gentle decline but an abrupt correction that leaves behind significant destruction of economic value and institutional trust. In this sense, artificial intelligence does not behave like an ephemeral trend but like a phenomenon with a speculative dynamic, fueled by a combination of hope, competition, and fear. What sets it apart from a simple fad is its potential to trigger macroeconomic and social consequences on a large scale, something no technological fad on its own has achieved in recent history.

1.3. The relationship between expectations, promises, and technical reality

The gap between expectations, promises, and technical reality is one of the most defining elements of the current artificial intelligence bubble. In no other field of recent history has there been such a wide distance between what is announced, what is believed possible, and what can actually be done with the technological resources available. This divergence is not only the result of natural enthusiasm in the face of a disruptive breakthrough, but of the way the media, corporate, and political narrative has turned AI into a symbol of unavoidable progress, projecting onto it expectations that far exceed its real capabilities.

Today, artificial intelligence is perceived as a system capable of thinking, reasoning, and creating autonomously, when in reality even the most advanced models such as large language models or multimodal generation systems lack semantic understanding, consciousness, intention, or independent reasoning ability. They operate through statistical correlations between data, not through conceptual understanding of the world. Yet this technical distinction, essential for understanding their limitations, rarely appears in public discourse. The consequence is that society’s perception of AI is deeply mythologized: it is associated with human intelligence when in fact it is simulated intelligence.

The contrast between promises and achievements is also evident in the economic sphere. Global consulting firms estimated that AI would generate an impact of more than 15 trillion dollars on global GDP by 2030, but recent reports from the International Monetary Fund and PwC indicate that less than 20 percent of companies have successfully incorporated generative AI models with a positive return on investment. Pilot projects are abundant, but real and sustainable implementation is far lower than what is publicly communicated. In addition, 80 percent of models deployed in corporate environments still depend on constant human supervision, manual reviews, and corrections of systematic errors, which contradicts the idea of full automation that dominates popular discourse.

Another critical aspect is the scale of energy and resource consumption compared to the value generated. Training a single large language model can consume more than 1,200 megawatt-hours of electricity, equivalent to the consumption of a medium-sized city for several days. The infrastructures required to sustain these systems demand multibillion-dollar investments and vast amounts of water for cooling. Yet the actual productivity obtained from these models, outside of content generation and automated customer service, remains marginal compared to the total cost of their deployment. This asymmetry between cost and performance is another indicator that the sector’s growth is driven by the narrative of potential rather than present profitability.

The problem is amplified by the way the tech industry communicates its progress. Large corporations tend to present each update or new version as a quantum leap toward artificial general intelligence, even though improvements are usually incremental. The press, searching for eye-catching headlines, amplifies these declarations, creating a public perception of exponential progress that does not correspond to technical reality. This cycle of expectations feeds itself: investors react to perception, governments push policies based on that perception, and companies compete to maintain media attention, leading them to promise more than they can deliver.

The distance between promises and reality is also seen in alleged cognitive breakthroughs. Language models can produce coherent texts but do not understand their content; image generation systems create complex visual representations without knowledge of context or artistic intention. In fields such as medicine or biology, progress is promising, but still depends on human verification and statistical validation. Nonetheless, the dominant narrative tends to present these achievements as milestones heralding imminent superintelligence.

At the political and social level, this expectation gap has direct consequences. Governments and international organizations have begun designing national AI strategies based on scenarios that assume a level of technological maturity that does not exist. Public investment programs are justified with inflated figures of potential productivity, and the education sector is accelerating the training of “AI experts” who, in many cases, lack practical experience with real models. This inflation of expectations results in an overproduction of discourse and an underproduction of tangible outcomes, increasing the pressure to justify increasingly large investments.

The result is an ecosystem where promises determine decisions more than evidence. Developers are forced to maintain an innovation pace that often does not respond to real needs, but to market pressure. Companies launch “AI-powered” products that barely incorporate automated functions, while the public associates any digital advancement with artificial intelligence. This environment creates an illusion of continuous progress, when in fact much of current development is aimed at sustaining the perception of progress.

The relationship between expectations, promises, and technical reality in the context of AI reveals a phenomenon in which discourse has surpassed technology. It is not artificial intelligence that is failing to fulfill its promise, but the economic, media, and political ecosystem that has turned that promise into a tool for speculation, positioning, and propaganda. Hype inflated with sensationalism. That is the psychological and structural foundation of every bubble: when narrative replaces reality, and faith in the future becomes a business model.

1.4. The disconnect between real economic value and stock market or media valuation

The disconnect between real economic value and stock market or media valuation is one of the clearest and most concerning indicators that artificial intelligence has crossed the threshold separating sustainable innovation from a speculative phenomenon. What began as a legitimate investment process in emerging technologies has turned into a dynamic of artificially inflated value, where perception and narrative weigh more than financial results, and where future expectations determine present prices that are completely detached from economic fundamentals.

In the past two years, the figures confirm this trend. According to data from Nasdaq and Bloomberg, the five largest technology companies directly linked to the development or exploitation of artificial intelligence (Nvidia, Microsoft, Alphabet, Amazon and Meta) currently concentrate more than 28 percent of the total value of the S&P 500 index. This level of concentration has no precedent, not even during the dot-com boom in 2000, when the weight of major tech companies did not exceed 19 percent. In the case of Nvidia, its market capitalization exceeded 2.8 trillion dollars in October 2025, despite annual revenues barely surpassing 61 billion, resulting in a price to earnings ratio above forty times its net profits, a level of overvaluation that is difficult to justify on industrial grounds.

The situation is repeated among startups and emerging companies in the AI ecosystem. Firms that have not yet achieved profitability or demonstrated a viable business model receive market valuations higher than established corporations in traditional productive sectors. OpenAI, for example, was valued at more than 150 billion dollars in private investment rounds in 2025, despite reporting estimated losses of 13 billion in the same fiscal year. Anthropic, Cohere and Mistral AI have each attracted more than 10 billion in capital, even before offering stable final products or having defined return metrics. This dynamic is reminiscent of the inflated valuations of the dot-com bubble, where perceived value depended more on the promise of the future than on real performance.

Media exposure amplifies this disconnect. Specialized and general press coverage of AI has grown exponentially since 2023, turning every incremental advance into a historic event and every new model version into a global turning point. Technology companies take advantage of this visibility to boost their valuation through the construction of an epic narrative: that of being pioneers in the most important revolution since electricity or the Internet. The media, in turn, feed this narrative because it generates traffic, clicks and public attention. A feedback loop is formed in which stock market value inflates not because of results, but because of the collective perception that we are witnessing an inevitable transformation with unlimited growth.

The problem is not only financial but structural. Inflated valuations distort capital allocation and divert investment away from productive sectors toward highly speculative technological projects. Sovereign funds, investment banks and pension funds have redirected a substantial portion of their portfolios toward AI, assuming that future growth will compensate for the current lack of profitability. This introduces a systemic risk: if the expected profits do not materialize, the market correction will affect not only the tech sector but the stability of global financial markets. Economic history shows that bubbles do not burst only due to lack of results, but because the system sustaining them loses the confidence that inflated them.

Another example of this disconnect is the contrast between market capitalization and real contribution to GDP or employment. While AI companies multiply their stock value, their impact on net employment and aggregate productivity remains limited. According to the International Labour Organization, the number of jobs created directly by the artificial intelligence sector represents less than 0.1 percent of the global total, and the increase in productivity resulting from industrial adoption remains around 1 percent. In contrast, the market value of companies in the sector has multiplied tenfold in three years. This gap between the real value contributed and the financial value assigned is a classic sign of speculative inflation.

The media discourse also helps consolidate a distorted vision of what constitutes success in the sector. Investment attraction is rewarded rather than profitability. Infrastructure expansion is celebrated rather than efficiency. And visibility is confused with impact. Companies compete to occupy headlines and to associate their brand with the concept of AI, even when their technical relationship with artificial intelligence is minimal or superficial. This has caused the term to become a synonym for innovation without requiring technical justification, generating an inflationary effect of confidence that distorts risk perception.

This disconnect between economic value and stock market valuation is also reflected in political narrative. Governments, attracted by the promise of technological leadership and media pressure, have begun allocating multimillion-dollar budgets to AI programs without clear impact indicators. The European Union, the United States, China, Japan and South Korea have announced combined public investment plans exceeding 500 billion dollars between 2024 and 2027. However, most of these programs lack auditing and evaluation mechanisms to measure the real effectiveness of these investments.

Artificial intelligence has thus become more a symbolic asset than an economic one: a banner of prestige and geopolitical power. Its value is no longer measured in profits or productivity, but in the perception of leadership and the narrative of the future. This disconnect between economy and myth, between capital and confidence, is the precise point at which the bubble ceases to be a possibility and becomes a reality.

1.5. The paradox of the “boom of the century”: a promised revolution at risk of imploding from excessive faith

The paradox of the so-called “boom of the century” lies in the fact that artificial intelligence, the very technology that has been proclaimed as the engine of a new era of progress, productivity, and global transformation, could be threatened precisely by the euphoria it has generated. Never before has a technological innovation concentrated so much capital, political attention, media coverage, and collective hope in such a short time. However, this very intensity —this excess of faith in its ability to solve all human, economic, and social problems— is creating the conditions for its own fragility. What began as a scientific revolution runs the risk of turning into a structural crisis if the pace of promises continues to outstrip that of results.

Artificial intelligence is living a historic moment comparable, in scale and expectations, to the invention of the combustion engine or the transistor. In just two years, global spending on AI projects and products has increased eightfold, while the term “ROI” (return on investment) is turning into science fiction for most companies, and the stock market valuations of the main firms in the sector have reached unprecedented levels. According to Goldman Sachs, total investment related to AI will exceed 500 billion dollars in 2026, driven by sovereign funds, venture capital, and public policies aimed at full digitalization. It is, without a doubt, the greatest mobilization of resources in modern history towards a single technology. But, paradoxically, that very scale also contains its vulnerability.

The central problem of this boom is not the technology itself, but the way in which society has loaded it with symbolism. AI has gone from being a tool for automation and analysis to becoming a modern myth: a promise of economic salvation, of superior intelligence, and of total efficiency. Companies and governments not only fund it; they revere it. It is presented as a force capable of replacing human thought, anticipating decisions, curing diseases, eliminating repetitive work, and even solving climate change. In this context, AI has ceased to be a scientific discipline and has been transformed into an ideological narrative about the future.

The paradox of the boom lies in the fact that the more AI is idealized, the further it drifts from its technical essence and the more dangerous its evolution becomes. Instead of consolidating itself as a strategic tool for progress, AI is being used as a symbol of status and institutional modernity. In the corporate sphere, companies of all sizes announce artificial intelligence projects without a clearly defined functional purpose, merely to align themselves with the trend of the moment and attract investment. In the political sphere, governments compete to demonstrate technological leadership, promising “revolutions” they cannot guarantee. In academia, thousands of studies and models are published that replicate the same results under different names, contributing to a saturated, redundant, and fragmented ecosystem.

Excessive faith also has practical consequences. By generating unattainable expectations, any slowdown or outcome that falls short of what was promised can trigger a generalized loss of confidence. Economic history has shown that bubbles do not collapse because of a lack of innovation, but because of the collapse of the faith that sustains them. The nineteenth-century railway bubble, the dot-com euphoria, or the 2008 mortgage crisis all share the same psychological pattern: the conviction that a new era had been reached in which market rules and economic prudence no longer applied. In the case of AI, the claim that “this time is different” resounds strongly, which is precisely what makes a repetition of history more likely.

Now more than ever there are economic interests at stake, and these take precedence over everything else. That is why, now more than ever, you will be told the same thing they have always said: “There is no bubble. Everything is fine.”

On top of this there is a fundamental contradiction: while artificial intelligence is presented as a force that will multiply global productivity, current data do not reflect a proportional increase. The “Global AI Outlook 2025” report by McKinsey indicates that, despite the sector’s growth, overall productivity in developed countries has barely risen by 1.2% since the mass adoption of generative AI. At the same time, energy, hardware, and model maintenance costs have soared to levels that threaten future profitability. In other words, and as I have already mentioned, the promised revolution is not yet producing an economic impact that justifies the scale of the investment, but it is generating a technological dependency that may become unsustainable.

This excessive faith also affects public perception. The narrative that AI “will do everything better” is creating a climate of cognitive and technological resignation, in which individuals and institutions begin to delegate critical judgment to systems that they themselves do not understand. This dependency, combined with media sensationalism, reinforces the central paradox of the boom: the more artificial intelligence promises, the weaker the human intelligence that supervises it becomes.

The “boom of the century” is, therefore, a double paradox. It reflects the greatest scientific advance of recent decades and, at the same time, mirrors our emotional vulnerability as a species. The obsession with achieving superintelligence has led us to overvalue what is artificial and undervalue what is human. History shows that every technological revolution feeds on faith, but it also shows that every unmeasured faith leads to excess. In the case of artificial intelligence, that excess takes the form of irrational investments, impossible expectations, and an almost religious trust in a future that does not yet exist. And when faith replaces reason, the fall is not only economic, but cultural.

2. Historical parallels with other bubbles and crises

To understand the magnitude of the phenomenon represented by the current artificial intelligence bubble, it is necessary to look back and examine the crises and bubbles that have shaped the economic and technological history of humanity. No bubble appears out of nowhere: all of them share psychological, financial, and social patterns that repeat with unsettling precision. In every era, a new promise of progress has served as fuel for collective euphoria, and on every occasion, the excessive faith placed in an idea, a technology, or a market has ultimately collapsed under the weight of its own expectations.

Studying historical parallels is not an exercise in economic nostalgia but a form of intellectual self-defense against the repetition of past mistakes. Humans tend to ignore warnings when they believe they are living through a unique and unrepeatable moment. However, every generation that has uttered the phrase “this time is different” has ended up facing the consequences of having mistaken innovation for infallibility. Acknowledging the lessons of the past does not mean denying the transformative potential of artificial intelligence, but accepting that even the most legitimate technological revolutions can be distorted by speculation, fear of being left behind, and the irrational pursuit of immediate profitability.

This is why it is essential to revisit the historical episodes that defined cycles of boom and collapse over the past centuries, from the Tulip Mania of the seventeenth century to the global financial crisis of 2008. In each of these moments, the same pattern repeated itself: a brilliant idea turned into a symbol of infinite wealth, a collective enthusiasm that clouded judgment, a credit expansion that seemed inexhaustible, and finally, the inevitable return to reality. Only by understanding how and why those crises occurred can we clearly analyze what is happening today and perhaps prevent artificial intelligence—the greatest promise of the twenty-first century—from becoming its greatest disappointment.

2.1. The Tulip Bubble (1637): the first speculative frenzy of modern history, driven by fashion and social prestige

The Tulip Bubble, also known as Tulip Mania, is considered the first documented episode of large-scale financial speculation in modern history. It emerged in the Netherlands at the beginning of the seventeenth century, at a time when international trade and economic growth had turned Amsterdam into one of the most prosperous financial centers in the world. In this context of abundance and optimism, a simple flower—the tulip—became a symbol of status, beauty, and wealth. What began as a fashion among the social elite eventually turned into a collective fever that swept up merchants, artisans, farmers, and bankers, culminating in a devastating collapse in 1637.

The origin of the phenomenon was aesthetic rather than economic. The tulip, introduced from the Ottoman Empire into Europe in the mid-sixteenth century, was an exotic flower, difficult to cultivate and displaying unusual colors. Some varieties had unique streaks and tonal patterns caused by a plant virus, making them extremely rare and highly coveted. Possessing rare tulips quickly became a symbol of social prestige and cultural refinement, a kind of class marker in Dutch society. But what began as an aristocratic trend turned into a collective obsession once the trade of bulbs became an opportunity for quick enrichment.

By 1634, the tulip market had ceased to operate as an exchange of real goods and had become a speculative futures market. Traders were no longer buying flowers but promises of flowers: contracts representing the right to acquire bulbs in the next season. This mechanism allowed buyers and sellers to trade claims on tulips that did not yet physically exist, a practice similar to modern financial derivatives. Prices multiplied irrationally. A single bulb of the most prized variety, the Semper Augustus, sold for more than 5,000 florins—an amount equivalent to the value of a house in Amsterdam or the lifetime annual salary of a master craftsman.

Enthusiasm spread rapidly. Bakers, carpenters, sailors, and farmers began participating in bulb trading, convinced that prices would never fall. Tulips became an abstract financial asset, completely disconnected from their intrinsic value. Newspapers and pamphlets of the time celebrated stories of individuals who had risen from poverty to wealth in a matter of weeks, further fueling the frenzy. The collective psychology of easy profit prevailed over economic reason, and the market expanded at a dizzying pace.

The turning point came in February 1637. At a public auction in Haarlem, the price of a lot of tulips failed to reach the expected figure and no one wanted to buy it. That small incident was enough to shatter general confidence. Within days, buyers disappeared and the value of bulbs collapsed by up to 95 percent. Thousands of contracts went unpaid, fortunes evaporated, and the local economy suffered a sudden contraction. The Dutch government attempted to intervene to stabilize the situation, but it was already too late. The bubble had burst.

The episode of Tulip Mania left a lesson that remains relevant almost four centuries later: the psychological dynamics of bubbles do not depend on the nature of the object, but on the perceived value attributed to it. Tulips were nothing more than flowers, but they became vehicles of speculation and symbols of status. The human desire for belonging, recognition, and immediate enrichment fueled the collective illusion that value could grow indefinitely. The same happens today with many emerging technologies: they are idealized, overstated, and elevated to the category of inevitable revolutions until reality shows that no growth can be sustained by faith and social prestige alone.

The Tulip Bubble was, ultimately, the first mirror of a behavior that humanity would repeat again and again: replacing real value with perceived value, and believing that prices can always rise simply because everyone else believes the same. A psychological pattern that today, four centuries later, reappears in a new, more sophisticated and technological form, but with the same underlying essence.

2.2. The railway bubble of the 1840s: promises of universal connection, massive investments without real profitability

The railway bubble of the 1840s represents one of the clearest examples of how a legitimate technological innovation can transform into a phenomenon of irrational speculation when expectations surpass economic reality. Railway expansion was undoubtedly one of the most transformative revolutions of modern history. It reshaped trade, shortened distances, and connected regions that had previously remained isolated. However, the euphoria that accompanied its growth in the United Kingdom and later throughout Europe and the United States evolved into a financial collapse that demonstrated how even the most useful technologies can generate bubbles when they become the object of unlimited promises and collective greed.

At the beginning of the nineteenth century, the railroad became the ultimate symbol of industrial progress. The invention of the steam locomotive by George Stephenson and the success of the Liverpool–Manchester line in 1830 showed that land transportation could be radically faster and more efficient than any previous method. The possibilities seemed endless: expansion of trade, national integration, job creation, and land revaluation. In just a few years, the so called Railway Mania took over the United Kingdom. Thousands of investors, from aristocrats to workers, began purchasing shares in railway companies convinced that the train would bring guaranteed prosperity to every town and every city.

Between 1835 and 1846, the British Parliament approved more than 9,500 miles of new railway lines, a volume that vastly exceeded the country’s technical and economic capacity. More than 800 railway companies were founded, many of them lacking experience, real planning, and in some cases, even lacking the intention to build anything at all. Newspapers published new investment opportunities every day, banks issued loans without sufficient guarantees, and engineers became public celebrities. By 1845, railway stocks already accounted for more than 40 percent of the total value of the London Stock Exchange.

But behind this apparent boom, signs of imbalance were emerging. Construction costs were enormous, demand forecasts were inflated, and many routes were planned based on political or speculative interests rather than logistical needs. Access to cheap credit encouraged the creation of redundant lines and routes that could never be profitable. The market began to saturate, and reality set in when the first financial statements revealed that profits were far lower than expected. As investors realized that many lines would never generate returns, enthusiasm quickly turned into panic.

The collapse erupted in 1846. Railway stocks began falling abruptly, dragging down banks, investment funds, and individual savers. Thousands of small investors lost their savings, and dozens of companies went bankrupt before laying a single rail. Parliament attempted to curb speculation by introducing new regulations requiring proof of technical and financial viability, but the damage was already done. In just two years, the total value of railway shares had been cut in half, triggering a recession that affected the entire British economy.

Despite this, the railroad survived as a technology and as an essential infrastructure. In the decades that followed, many of the lines built during the mania were consolidated, forming the basis of the modern railway system. However, the episode left a crucial lesson: a technology can be revolutionary while also becoming the center of a speculative bubble. The real usefulness of an innovation does not protect it from speculation, and collective enthusiasm can distort even the most solid advances.

The railway bubble of the 1840s showed that the promise of universal connection, at that time represented by iron rails, can become a trap when investment is driven by the desire to belong to the new era rather than by a rational analysis of expected returns. The parallel with artificial intelligence is evident. Today, as then, we hear of an invisible network that will connect all of humanity, the elimination of barriers, the total transformation of the economy, and the rise of new industries. And just like in 1840, capital flows with the conviction that nothing can go wrong because progress seems inevitable. But as history teaches, when faith replaces logic, even the strongest tracks can lead directly to derailment.

2.3. The bubble of the 1920s (the 1929 crash): technological euphoria driven by electricity, automobiles, and easy credit

The bubble of the Roaring Twenties, which culminated in the collapse of Wall Street in 1929, was the result of an explosive combination of technological innovation, social optimism, and blind faith in perpetual growth. It was a period marked by advances that radically transformed everyday life: the electrification of cities, the expansion of the automobile, radio, cinema, and the first household appliances created a widespread sense that progress no longer had limits. For the first time in history, technology became not only a tool for development but also a cultural symbol of modernity and success. Society believed it had entered a definitive era of prosperity, and that collective feeling fueled an economic euphoria that ultimately became one of the greatest crises in modern history.

The post–World War I context was decisive. The United States emerged as a global industrial and financial power, while Europe was immersed in reconstruction. Mass production, driven by the Fordist model, reduced manufacturing costs and multiplied middle-class access to goods once reserved for the elite. Automobiles, refrigerators, radios, and telephones began to fill households, shaping a new economy of desire and credit. Banks offered loans to individuals for purchasing products and to companies to finance their expansion. Debt became an everyday instrument, and the promise of endless growth fueled stock market speculation.

Between 1924 and 1929, the Dow Jones Industrial Average more than quadrupled, rising from 100 to 381 points. Shares of the technological companies of the era, such as General Electric, RCA (Radio Corporation of America), and Ford Motor Company, soared, reflecting not only actual profits but also the belief that the future would be an infinite extension of the present. The press celebrated the rise of the markets, universities taught that the new economy was invulnerable, and ordinary citizens invested their savings or took out loans to participate in the so called American miracle. It is estimated that more than four million Americans invested in the stock market, many of them without financial knowledge, convinced that the market could not fall.

Easy credit was the fuel of that bubble. Stockbrokers offered the possibility of buying shares on margin, meaning paying only a fraction of their value with one’s own money and the rest with borrowed funds. When share prices went up, profits seemed unlimited; but when they fell, losses multiplied just as quickly. This system of financial leverage created a fictitious debt-based economy in which the money invested far exceeded the money that actually existed.

Technological euphoria reinforced the collective illusion. Massive electrification, industrial production, and the rise of communications created the feeling of living through a revolution comparable to the discovery of fire or the invention of the wheel. The idea that humanity had entered an era of perpetual progress became the economic dogma of the decade. Technological advancement was confused with guaranteed prosperity. Technology, for the first time, was perceived as infallible, and financial markets as its inevitable reflection.

However, by mid-1929 warning signs began to appear. Agricultural prices fell, consumption started to slow, and companies accumulated unsold inventories. The real economy could no longer sustain the stock market’s growth. On October 24, 1929, known as Black Thursday, the New York Stock Exchange collapsed. In just a few days, more than 30 billion dollars evaporated, a figure equivalent to the annual GDP of the United States at the time. The ensuing chain of bank panics, corporate bankruptcies, and mass unemployment triggered the Great Depression, a period of unprecedented global economic contraction.

The lesson from that bubble remains relevant: technological progress, by itself, does not guarantee stability or prosperity. Faith in technology as an infallible engine of growth can become a trap when it is not accompanied by economic prudence, regulation, and financial education. In the 1920s, electricity, the automobile, and radio were the symbolic equivalents of today’s artificial intelligence: real innovations that transformed society but were also used as justification for an economic expansion based more on promise than on productivity.

History shows that every technological revolution generates a moment of euphoria in which the collective imagination surpasses the limits of reality. In 1929, the belief in eternal progress collapsed overnight, reminding the world that no technology, no matter how transformative, can escape the basic laws of economics or the human psychology of excess. And just as electricity and the combustion engine were symbols of the optimism of their time, artificial intelligence occupies that place today: a promise that, if not managed with balance and realism, may end up following the same fate as that brilliant and tragic mirage of modernity.

2.4. The dot-com bubble (1999–2001): the closest precedent to today’s AI phenomenon; excessive promises, companies without business models, and a blind faith in the digital future

The dot-com bubble, which took place between 1999 and 2001, was one of the most representative episodes of modern technological speculation and, without question, the closest precedent to the current artificial intelligence bubble. In both cases, a genuine innovation — the Internet then, AI now — was accompanied by an explosion of expectations that turned a technical revolution into a financial frenzy. The phenomenon was based on a simple but powerful idea: that the world was on the verge of changing forever, that the new digital economy would replace the old one, and that any company associated with that future deserved to be funded without questioning its present profitability.

At the end of the 1990s, Internet access was expanding at a dizzying pace. Global connectivity, e-commerce, and digital communication promised to redefine every productive sector. The media and financial analysts fueled a narrative of infinite prosperity. People spoke of “the new economy,” a system supposedly freed from the traditional rules of the market, where companies did not need profits to be valuable. What mattered was not profitability, but “exponential growth,” “disruption,” and “the ability to acquire users.” This discourse became a dogma that justified any valuation, no matter how absurd, as long as it was associated with a web domain and a digital dream.

Between 1998 and 2000, more than 5,000 new tech companies were registered in the United States. Many of them were created without a clear business model, without revenue, without a functional product, and even without a sustainability strategy. Companies like Pets.com, eToys, Webvan, and Kozmo became emblematic examples of the era: firms that spent millions on marketing and expansion in the hope of dominating a future market, but that never managed to generate real profits. The mantra was “grow first, make money later.” Investors were guided more by enthusiasm and fashion than by rationality.

The flow of capital was overwhelming. Initial public offerings (IPOs) multiplied, and the stock prices of newly created companies skyrocketed within hours of going public. In 1999, the Nasdaq Composite — the index that concentrated most tech companies — tripled in value, reaching 5,048 points in March 2000. It was unprecedented growth, sustained more by expectation than by productivity. Institutional investors, banks, and venture capital funds competed to invest in any company that had “.com” in its name, convinced that the Internet would eliminate the limits of the physical market and create a new form of unlimited wealth.

However, beneath this euphoria lay a structural problem: most companies were not generating profits and depended entirely on external financing. Spending on advertising, servers, and digital infrastructure grew faster than revenue. When interest rates began to rise and investors demanded concrete results, the illusion collapsed. Between March 2000 and October 2002, the Nasdaq lost almost 80 percent of its value. Thousands of companies went bankrupt, and hundreds of thousands of people lost their savings and jobs. Pets.com, which had spent more than 100 million dollars on marketing in less than a year, closed just nine months after going public.

The crash had deep consequences. Confidence in the tech sector plummeted, and investments dropped drastically during the following years. However, the infrastructure built during that period — fiber-optic networks, data centers, payment systems, and digital culture — became the foundation for the technological resurgence of the next decade. From the ashes of the bubble emerged giants like Google, Amazon, and eBay, which managed to turn the initial chaos into sustainable models. The lesson was clear: technology can be revolutionary, but the market cannot be sustained on faith alone or on growth without purpose.

The similarity between the dot-com bubble and today’s AI fever is evident. In both cases, the dominant narrative revolves around the idea that we are witnessing the birth of a new economic era. Today, just as then, people invest in promises rather than products, value speed over robustness, and reward media visibility over technical sustainability. Startups that have barely launched prototypes are valued in the billions; traditional companies reinvent themselves with the “AI” label to attract capital, and the media present every incremental advance as the prelude to an imminent superintelligence.

In 1999 people said “everything will change with the Internet”; in 2025 we hear that “everything will change with artificial intelligence.” In both cases, the statement is partially true: both technologies are transformative. But the mistake lies in confusing transformation with redemption, progress with infallibility. The history of the dot-com bubble teaches that excessive faith, easy credit, and an obsession with growth without substance can turn a real revolution into an economic mirage. If artificial intelligence continues to fuel the same dynamics — investments without returns, inflated expectations, and dependence on media hype — its outcome is likely to repeat, on a larger and more complex scale, the same pattern that marked the end of the first digital dream of the twenty-first century.

2.5. The 2008 financial crisis: speculation on intangible assets (derivatives, mortgages) that seemed safe until they collapsed

The global financial crisis of 2008 was the most devastating economic collapse since 1929 and stands as a paradigmatic example of how speculation on intangible assets can create an illusion of stability and wealth that is in fact built on fragile foundations. In this case, the object of the bubble was not a flower, a railway, or a technology company, but a financial instrument: mortgage-backed derivatives. What was theoretically a sophisticated mechanism for risk management became a machinery of uncontrolled speculation that ultimately brought down the international financial system and dragged millions of people into the loss of their homes, jobs, and savings.

During the first half of the 2000s, the United States housing market experienced unprecedented growth. Historically low interest rates and the expansion of credit driven by banks encouraged millions of families to purchase homes, even without real repayment capacity. Financial institutions, instead of assuming the risk of those loans, began packaging them into instruments known as mortgage-backed securities (MBS) and collateralized debt obligations (CDO). These products were sold on international markets to funds, banks, and insurance companies, which considered them safe investments because they were backed by tangible assets, the houses themselves, and because of the high credit ratings given by rating agencies.

The system seemed perfect: banks issued more mortgages, housing prices rose, investors profited, and credit continued to flow. In reality, a parallel economy had been created based on recycled debt and blind trust in financial engineering. Derivatives made it possible to transform high-risk loans into seemingly stable assets. Financial institutions assumed that the housing market could never fall simultaneously across the entire country, a premise that would prove disastrous. Faith in mathematical sophistication replaced prudence and economic analysis.

Between 2003 and 2007, the value of securitized mortgage assets in the United States grew from 2.5 to more than 8 trillion dollars. Investment banks achieved record profits, executives received multimillion-dollar bonuses, and the housing market soared. Easy credit and speculation pushed housing prices to artificially high levels: by 2006, the median home value in the United States was 85 percent higher than in 2000. Warning signs, such as excessive household debt, the rise of high-risk subprime mortgages, and the lack of transparency in financial products, were ignored or minimized under the argument that risk models were under control.

The bubble began to deflate in 2007, when interest rates rose and millions of homeowners started defaulting on their mortgages. Banks discovered that the assets they had considered safe were, in fact, uncollectible. Within months, the world’s largest financial institutions were trapped in a network of interconnected debts. In September 2008, the collapse of Lehman Brothers, one of the oldest and most prestigious firms on Wall Street, marked the point of no return. Panic spread through global markets, stock exchanges plummeted, and the international financial system stood at the brink of total collapse.

The impact was catastrophic. Millions of people lost their homes and jobs; governments were forced to bail out banks with public funds; and trust in the financial system suffered a blow from which it would take years to recover. According to the International Monetary Fund, total losses resulting from the crisis exceeded 22 trillion dollars in destroyed assets and more than 60 million direct and indirect jobs. The episode exposed the disconnect between the real economy, based on production, employment, and consumption, and the speculative economy, based on expectations, debt, and opaque financial algorithms.

The parallel with the current phenomenon of artificial intelligence is evident. In 2008, banks believed they had found an infallible formula to create value from intangible products and dispersed risks. Today, big tech companies and investment funds repeat a similar dynamic by financing AI projects whose profitability depends on future expectations rather than present results. Just as financial derivatives were labeled safe assets due to their apparent sophistication, AI models are presented as inevitable and transformative, even though much of their value remains theoretical.

The 2008 crisis showed that complexity does not equal safety and that technical opacity can be as dangerous as ignorance. Back then, financial algorithms replaced human analysis; today, artificial intelligence algorithms promise to replace decision-making itself. In both cases, the risk is the same: the blind delegation of responsibility and the belief that technology, by its mere existence, eliminates human error. History has shown that the true bubble does not lie in the assets but in the collective mindset that chooses to believe the future can be built without limits or consequences.

2.6. How every bubble shares the same psychological pattern: the promise of a new era, the mass participation of actors without deep knowledge, and the belief that “this time is different”

All bubbles, regardless of their era, context, or the type of asset that triggered them, share the same psychological pattern deeply rooted in human nature. Behind every speculative boom, whether tulips in the seventeenth century, railroads in the nineteenth, tech stocks in the twentieth, or artificial intelligence in the twenty first, the same impulse is always present: the promise of a new era that will forever change the fate of humanity. This narrative, which mixes hope, greed, and an almost religious faith in progress, is the invisible engine that drives collective euphoria. In every cycle, the words change, the protagonists vary, but the emotional structure remains identical.

The first element of that pattern is the promise of a new era. Every bubble begins with an innovation or structural shift that sparks genuine enthusiasm. In the Tulip Mania, it was the exotic beauty turned into a status symbol; in the railroad era, the possibility of connecting entire countries through steel; in the twenties, electrification and the automobile; in the dot com years, the digital revolution and global access to information; and today, artificial intelligence with its apparent capacity to surpass the limits of human thought. In all cases, the narrative of absolute progress prevails over economic prudence. A conviction takes hold that the world is facing an irreversible transformation, that the future has already arrived, and that anyone who does not participate will be left outside history.

The second common element is the massive entry of actors without deep knowledge. As enthusiasm spreads, the complexity of the phenomenon becomes simplified and turns into an opportunity open to everyone. In seventeenth century Holland, artisans and farmers invested in bulbs without understanding their biological nature or the commercial risk; in Victorian England, small savers bought railroad shares convinced that simply owning “a track” was enough to become rich; in the twenties, housewives, workers, and employees invested their savings in the stock market without understanding market volatility; and during the dot com fever, millions of retail investors bet on digital startups with no product and no profits. Today something similar is happening with artificial intelligence: governments, companies, and individuals invest resources, time, and credibility in projects whose technical functioning they do not understand, driven by the fear of being left behind and the illusion of taking part in historic change.

The third component of the psychological pattern is the belief that “this time is different.” It is the most repeated and most dangerous phrase in economic history. Every generation, convinced of its moral and technological superiority, believes it has overcome the mistakes of the past. The Dutch in the seventeenth century thought their commercial prosperity made a crisis impossible; the British in the nineteenth century believed industrialization guaranteed infinite growth; Americans in the twenties claimed their new financial system was infallible; and Silicon Valley entrepreneurs in 1999 insisted that the Internet had abolished the laws of the traditional market. Today, the most enthusiastic defenders of artificial intelligence claim that its potential is so extraordinary that no collapse can occur, because the technology feeds back on itself, learns, and self corrects. But the thinking is the same: the illusion of immunity to the economic cycle and the belief that the future is no longer subject to human error.

Deep psychological factors also feed into this cycle. The desire for quick wealth, the fear of exclusion, and the social pressure not to be “left out” fuel the speculative mass. Every bubble contains an element of emotional contagion, the idea that everyone is winning and that anyone who hesitates simply does not understand the change. This collective phenomenon generates a kind of critical anesthesia that eliminates prudence and replaces logic with imitation. What matters ceases to be the value of the asset and becomes the social validation provided by participating in the euphoria. Today this dynamic is amplified by social networks, the media, and the immediacy of information, which allow expectations and disappointments to spread at unprecedented speed.

The pattern always ends the same way: reality eventually imposes itself. When prices stop reflecting fundamentals, when promises fail to materialize, and when returns never arrive, confidence breaks. And when confidence disappears, the entire system collapses. What is presented as the beginning of a new era ends up becoming a lesson in collective humility.

As you can see for yourself now, history shows that bubbles are not isolated economic errors but recurring manifestations of human psychology in the face of progress. Each one is born from a combination of real innovation and unrealistic expectations. Each one promises definitive transformation, and each one ends up reminding us of the same thing: that progress is not linear, that technology does not eliminate risk, and that no revolution, no matter how brilliant it appears, is exempt from repeating the same mistakes when enthusiasm replaces judgment. The artificial intelligence bubble is not an exception but the most sophisticated version of a pattern as old as the human desire to believe that this time, truly, everything will be different.

3. The elements that fuel the current AI bubble

Understanding the nature of the artificial intelligence bubble requires examining the factors that feed and sustain it. Although the phenomenon appears new because of its technological dimension and global reach, its internal structure follows the same mechanisms that have driven all previous bubbles: excessive expectations, capital oversupply, media narratives, and a collective belief in unlimited progress. The fundamental difference is that this time the scale is planetary, and the object of speculation is not a physical good nor a financial asset, but an abstract promise: the idea that artificial intelligence will completely transform the economy, science, politics, and human life. And even if it might, that does not mean it will. Everything depends on the path taken to reach the objective, and the error and major problem here is the path that has been chosen. This is why it is highly likely that humanity itself, through its hypocrisy, ignorance, and mental illiteracy, will destroy the only chance it has to survive its own future of self-destruction and the extermination of its own species, as well as the planet that gives it life and a home.

The current bubble is not built upon a single economic mistake, but on a complex framework where corporate interests, political ambitions, media disinformation, and technical ignorance converge. Large technology corporations compete to dominate the narrative of the “new intelligent era,” investors seek quick returns, governments want to display digital leadership, and the public enthusiastically embraces a story that promises efficiency, knowledge, and power. All of this creates an ecosystem where growth is perceived as inevitable, criticism as backwardness, and caution as failure.

Examining the elements that fuel this bubble makes it possible to understand why the discourse on artificial intelligence has shifted from innovation to dogma. At this point, I want to outline the structural, economic, and psychological factors that are inflating its value beyond its technical reality, creating an imbalance that, if not corrected, could reproduce the same consequences that in the past turned enthusiasm into crisis.

3.1. The Overvaluation of Companies Without Profitable Models

One of the most visible pillars sustaining the artificial intelligence bubble is the massive overvaluation of companies that lack profitable or sustainable business models. This dynamic, driven by excess liquidity, competition among investment funds, and the media frenzy surrounding AI, has created a financial environment in which the market value of many companies bears little relation to their actual ability to generate profits. What is rewarded is not profitability or stability, but the promise of the future: the expectation that, at some point, the artificial intelligence revolution will produce returns proportional to the astronomical figures currently assigned to it.

The most emblematic case of this trend is OpenAI, whose private valuation in 2025 exceeds 150 billion dollars, despite posting estimated annual losses of more than 13 billion, according to sources from Bloomberg and The Wall Street Journal. These figures are comparable to those of established corporations such as General Motors or Siemens, companies that employ hundreds of thousands of people and produce tangible goods on a global scale. Other companies in the sector, such as Anthropic, Cohere, or Mistral AI, have reached valuations above 10 billion dollars in funding rounds, even though none of them has yet demonstrated sustained profitability or consolidated a predictable revenue stream. All of it based solely on a promise, the promise that AI is the sacred cure to all our problems.

The problem does not lie in investing in innovation, which is necessary and legitimate, but in the growing disconnect between the real value created and the financial value attributed. Investors are not buying products or results; they are buying expectations. In financial reports and presentations to shareholders, the emphasis is not on net profits or operating margins, but on growth projections, the number of registered users, or the volume of requests processed per second. Narrative has replaced the balance sheet. The same logic that inflated the dot-com bubble in the late nineties—grow first, monetize later—now repeats itself under new terminology: scale fast, dominate the market, capture data, or train the largest model.

Institutional investors, including sovereign wealth funds and major venture capital firms, fuel this cycle with an attitude that mixes greed and fear. On one hand, the greed to take part in the next industrial revolution; on the other, the fear of being left out of the technological wave that promises to define the future. This collective psychology, known as FOMO (fear of missing out), pushes large capital players to invest in unproven projects, guided more by the narrative of change than by economic viability. Consequently, billions of dollars flow into companies whose revenue structure depends almost entirely on external financing, without a clear plan for sustainability.

The result is an artificially inflated ecosystem in which valuations multiply due to the market’s mirror effect. When a leading company raises its valuation, others are automatically revalued to avoid falling behind in comparison. This phenomenon creates a vicious cycle: valuation attracts investment, investment increases valuation, and public perception reinforces the illusion of success. Meanwhile, financial fundamentals remain stagnant or even in decline. In many cases, revenue comes not from the sale of final products or services, but from subsidies, tax credits, or licensing agreements for their models to other companies that are also unprofitable.

At the same time, dependence on large tech corporations such as Microsoft, Google, Amazon, or Meta distorts the sector’s metrics even further. These companies, with multibillion-dollar margins from established businesses, subsidize their artificial intelligence divisions to support their narrative of global leadership. In this way, the apparent results of the sector are artificially inflated: losses are disguised as strategic investment, and operating costs are interpreted as innovation spending. In accounting terms, AI functions more as a marketing department than as a standalone industry.

The risk of this model is clear. When investment flows slow down or results fail to meet projections, the market will correct itself with the same speed at which it grew. Companies without profits, without stable customers, and without tangible competitive advantages will be exposed, and their value will evaporate as quickly as it was created. This scenario is not hypothetical; it has already begun to materialize. In 2025, several generative AI startups, particularly in the United States and Europe, initiated mass layoffs after exhausting their funding and discovering that their free user base did not translate into real revenue.

The overvaluation of unprofitable companies is not just a financial problem; it is also a cultural symptom. It represents a mindset that confuses innovation with invulnerability, and vision with faith. In this context, artificial intelligence has shifted from being a scientific discipline to becoming an instrument of narrative speculation. The price of the future is being paid today, with money that does not exist and with results that have not yet arrived. And as history has shown again and again, no market can sustain itself indefinitely on an unfulfilled promise.

3.2. The race to attract investment without solid products or real returns

The race to attract investment in the artificial intelligence sector has become one of the most intense and accelerated economic phenomena in recent history. What should have been a competition driven by technological innovation has instead turned into a contest to capture capital, media attention, and symbolic legitimacy. The result is an ecosystem dominated by the urgency to secure funding rather than the need to build solid, profitable, or sustainable products. This dynamic not only distorts the real development of the technology but also multiplies the risks of structural instability, repeating the same speculative patterns that preceded other major technological bubbles.

Over the past three years, the flow of capital into the artificial intelligence sector has reached unprecedented levels. According to data from PitchBook and CB Insights, more than 180 billion dollars were invested in AI startups between 2023 and 2025, most of them in early development stages and without proven business models. This amount exceeds the total invested in biotechnology, robotics, and fintech combined during the same period. However, the most significant aspect is not the volume but the nature of the investments: the majority is directed toward promises of the future, not operational products. The rhetoric of “disruption” and “global transformation” has replaced viability analyses, generating an economy of expectations in which value is built from narratives rather than results.

Funding rounds have become a measure of success in themselves. An AI company that manages to close a round worth hundreds of millions does not need to demonstrate revenue, customers, or real use cases; it only needs to convince investors that it possesses “the most advanced algorithm” or “the most efficient model.” Confidence is based on perception, not facts. In many cases, the product does not even exist, or it is in an experimental phase with no immediate commercial application. Nevertheless, valuations reach billions because funds compete to avoid being left out of the “next unicorn.” This behavior recalls the dot-com bubble, when simply owning a domain name was enough to access massive capital. Today, it is enough to include “AI” in the company name.

The corporate language of this new frenzy is almost identical to that of previous speculative cycles. Terms such as “exponential scale,” “continuous learning,” “cognitive automation,” or “artificial general intelligence” are used as narrative anchors to attract funds without providing verifiable data. Investor presentations focus on projected growth charts and the number of downloads or registered users, but systematically omit the real operating costs, the dependence on third-party infrastructure, or the absence of a clear revenue model. In practice, many startups survive thanks to a constant flow of external funding that is used to maintain the appearance of growth rather than to generate real value.

Another factor intensifying this race is the role of venture capital and sovereign wealth funds. Cheap money and global liquidity, combined with pressure to show rapid returns, have encouraged aggressive investments in high-risk projects. Funds that in previous decades invested in biotechnology or clean energy have redirected their resources toward AI, attracted by the possibility of multiplying their capital in a short time. But the lack of technical criteria in project evaluation has created a cascading effect: investments are made not in what is understood, but in what others are already financing. This imitation behavior, or bandwagon effect, turns venture capital into speculative capital, feeding a cycle in which investments are justified by the existence of previous investments, not by the intrinsic value of the project.

The result of this dynamic is a market where the narrative of growth replaces the reality of development. The most visible companies are forced to maintain an image of constant expansion to continue attracting capital, which obliges them to allocate enormous resources to marketing, public relations, and strategic alliances, at the expense of research and technical refinement. In some cases, scientific teams are reduced in order to prioritize commercial departments, and technological advances are announced before being verified, creating an illusion of continuous progress. In this way, the AI economy begins to resemble an attention economy more than an innovation industry.

The consequence is that the market inflates at a rate that cannot be sustained. According to a 2025 Deloitte report, 70 percent of AI startups founded between 2020 and 2024 generate no profits and depend entirely on successive funding rounds to remain operational. In addition, more than 60 percent of institutional investors admit that they do not fully understand the technical fundamentals of the companies they invest in. This creates a situation of vulnerability similar to the 2008 financial crisis, where complex financial products were acquired without a real understanding of their structure or risks.

The race to attract investment in artificial intelligence is, ultimately, an expression of modern faith in infinite growth. Each funding cycle promises to be closer to the “inflection point” that will justify current losses, but that point never arrives. Companies continue to increase their spending, investors continue waiting for returns, and the market continues betting that the future will save the present. This model, based on speculation rather than productivity, can be sustained as long as capital flows and the narrative remains intact. But when confidence runs out, the AI market will discover that the money invested did not buy innovation, but illusion.

3.3. The media use of the term “AI” for any software with basic automation

One of the most distorting elements in the current artificial intelligence bubble is the indiscriminate use of the term “AI” as a media and commercial tool. The concept, originally reserved to describe systems capable of performing tasks that require reasoning, learning, or understanding, has been degraded to the point of becoming an advertising label applied to practically any software with automated functions. This semantic misuse, driven by business interests and amplified by the media, has created an environment of massive confusion in which the boundaries between what is truly artificial intelligence and what is simply automation or advanced programming become blurred.

In practice, a large portion of the products and services presented today as “AI powered” do not meet the technical criteria that define that category. Recommendation systems, virtual assistants with predefined answers, search algorithms, automation macros, or simple statistical analysis processes are routinely marketed as artificial intelligence. Companies in sectors as diverse as banking, marketing, e-commerce, or human resources use the term as a synonym for innovation, even when their underlying technology does not include machine learning or real cognitive adaptation. The word “AI” has therefore become a positioning tool, a seal of modernity that generates trust and attracts investment, regardless of the technical accuracy of its application.

The media, in turn, play a crucial role in propagating this confusion. The constant pursuit of eye-catching headlines and the desire to participate in the narrative of progress have led to almost any technological advancement being labeled as “artificial intelligence”. A new search system, a photo editing tool, or a corporate chatbot are presented using the same language used to describe deep learning systems or multimodal generative models. This discursive uniformity produces an inflationary effect on the concept: the more the term is used, the less meaning it retains. The result is an illusion of omnipresence that makes the public believe that artificial intelligence already dominates every aspect of daily life, when in reality its technical penetration is much more limited.

Companies have learned to exploit this media effect to gain competitive advantages. Incorporating the “AI” label into brand communication or press releases can multiply investor interest, increase media coverage, and improve perceived product value. According to a study by consulting firm Gartner, startups that include the word “AI” in their name or description achieve, on average, 35 percent more initial funding than those that do not, even when operating in sectors where artificial intelligence has almost no direct application. This phenomenon reflects how perception has replaced evidence, and how semantics have become a tool for speculation.

The problem worsens when public institutions and policymakers adopt the same narrative without critical analysis. Governments, universities, and international organizations promote digitalization programs under the umbrella of “artificial intelligence”, allocating resources to projects that in many cases go no further than conventional automation. This not only disperses funds and dilutes the effectiveness of technological policies, but also contributes to creating a false sense of national progress in innovation. What should be a technological development strategy becomes a communication strategy.

This terminological inflation has direct consequences on society’s perception of AI. The public, constantly bombarded by headlines and advertisements, assumes that artificial intelligence is an all-powerful entity capable of understanding, deciding, and creating by itself. This distorted perception fuels both enthusiasm and fear, two emotions that together sustain the current bubble. The promise that AI will solve all problems coexists with the fear that it will eventually replace humans, a dual emotional landscape that maintains interest and capital flow without the need for concrete results.

The media use of the term “AI” reflects a deeper cultural phenomenon: the tendency of modern societies to turn technology into narrative. Artificial intelligence has ceased to be a scientific discipline and has become a symbol. Saying “AI” no longer means describing a technical capability but invoking an idea of power, progress, and control. And in that process, the term has lost precision but gained profitability. This indiscriminate use, although useful for selling and attracting investment, erodes the field’s credibility and generates a mirage of progress that will sooner or later collide with reality. Because when the label no longer matches the content, the market does not take long to adjust the value, and what is sold today as artificial intelligence risks being remembered as the most expensive automation in history.

3.4. Marketing and the messianic narrative: “AI will save the world”, “AI will make us immortal”, “AI will replace everything”

One of the most powerful engines of the current artificial intelligence bubble is the messianic narrative that has wrapped itself around this technology, a narrative carefully constructed and amplified by major tech corporations, the media, and an army of media “experts” who have turned AI into a kind of universal salvation promise. In the dominant discourse, artificial intelligence is not presented as a tool but as a redeeming entity capable of solving the world’s structural problems, transcending human limitations, and redefining the very concept of existence. Phrases like “AI will save the world”, “AI will make us immortal”, or “AI will replace everything” have shifted from visionary slogans to emotional truths within the collective imagination.

This phenomenon is neither accidental nor spontaneous. It is the result of a global marketing strategy that combines technological storytelling, philosophical aspiration, and corporate propaganda. Leading artificial intelligence companies have understood that selling technology is not enough; what truly mobilizes capital, attention, and loyalty is selling hope. Consequently, AI is presented as the key to a new era: an era without errors, without disease, without manual work, and eventually without death. In this context, engineers become prophets, CEOs become visionaries, and research labs become temples of the future. Technology ceases to be a tool and becomes a technoscientific faith.

The idea that artificial intelligence “will save the world” has spread like a modern dogma. It is claimed that it will solve climate change through energy optimization, end hunger thanks to agricultural automation, guarantee justice through impartial algorithms, and cure all diseases with predictive diagnostics. Each of these promises contains a real technical basis but is amplified until it reaches utopian dimensions. The complexity of social, economic, and ethical problems disappears from the narrative. AI, a tool designed by humans and trained on human data, is presented as a morally neutral and omnipotent force, a new “rational order” that will correct the failures of the species.

Added to this redemptive vision is an even more ambitious one: the promise of digital immortality. Some of the sector’s leading spokespersons, such as Ray Kurzweil or Sam Altman, have popularized the idea that advances in artificial intelligence and biotechnology will enable the fusion of mind and machine, the transfer of human consciousness to digital substrates, and the indefinite extension of life. This discourse, although lacking verifiable scientific support, has captured the imagination of millions and served as a powerful magnet for attracting private investment. The possibility of defeating death, humanity’s oldest aspiration, has become a sales argument. And like any effective myth, it does not need to be proven; it only needs to be believed.

The third great messianic promise is total replacement: the idea that artificial intelligence will replace all forms of human labor, creativity, and thought, liberating humanity from effort. In this vision, automation not only improves productivity but redefines civilization. Intelligent systems will write, design, compose, diagnose, legislate, and govern better than we can. Human beings, freed from mechanical and intellectual tasks, will be able to “think”, “create”, or simply “live”. However, this narrative omits an uncomfortable reality: replacement does not imply liberation, but displacement. Millions of workers, professionals, and creatives are already experiencing the consequences of this automation, while promises of collective prosperity remain, for now, theoretical.

The messianic rhetoric of artificial intelligence fulfills a precise economic and political function: maintaining the illusion of inevitability. If AI is the future, resisting it is equivalent to opposing progress. This discourse inhibits rational criticism and neutralizes ethical debate because it turns technology into destiny. Citizens cease to be active agents and become spectators of a supposedly inevitable transformation. The idea that “AI will replace everything” eliminates individual, institutional, and corporate responsibility, shifting focus away from human decisions and toward a technological abstraction presented as autonomous and morally neutral.

From a psychological perspective, this narrative appeals to the ancestral desire for transcendence. Confronted with the uncertainty of the future, humans seek absolute certainties, and technology offers a form of faith compatible with modern rationality. In the nineteenth century, the railroad was a symbol of civilization; in the twentieth, electricity and aviation; in the twenty-first, artificial intelligence occupies that role. But the difference is that now the object of worship not only promises to transform the world but also to redeem humanity from itself.

This technological mythification not only fuels the bubble but also makes it resistant to doubt. Questioning AI’s promises is interpreted as an act of ignorance or pessimism, while believing in them is associated with vision and leadership. The result is an ideological ecosystem where faith replaces evidence, and where the promise of a perfect future justifies any present excess. The risk is that, like all previous messianic narratives, the myth of technological salvation will eventually collapse under the weight of its own grandiosity. When the promises fail to materialize at the expected pace, the disillusionment will be proportional to the faith invested. And at that moment, what today is presented as the dawn of a new era may reveal itself as another human story of excess, belief, and downfall.

3.5. The overdependence on venture capital, sovereign funds, and governments

The expansion of the artificial intelligence sector over the past decade has been disproportionately driven by external financing from venture capital, sovereign wealth funds, and government programs. This massive flow of capital has enabled an unprecedented acceleration in the development of models, infrastructures, and startups, but at the same time it has created a structure of dependence that threatens the stability of the ecosystem itself. Instead of being built on the basis of sustainable revenue and real demand, much of today’s AI growth is supported by speculative capital injections that seek quick returns and visibility rather than economic or scientific impact.

Venture capital has become the main fuel for this expansion. Large funds, especially in the United States, Europe, and Asia, have transformed artificial intelligence into the top destination in their portfolios, displacing investments from traditional sectors such as biotechnology, clean energy, or robotics. According to data from CB Insights, between 2023 and 2025, 46 percent of total global venture capital investment went to projects directly or indirectly linked to AI. The logic behind this concentration is the search for the next technological giant: funds do not seek stability, but to multiply their initial investment in a short period of time, betting that one or two companies out of every hundred will reach a billion-dollar valuation. Although profitable in early phases, this model creates an artificially inflated ecosystem where the survival of companies depends less on their results than on their ability to continue attracting funding rounds.

Added to this dynamic is the participation of sovereign wealth funds, financial instruments controlled by governments, especially in the Middle East, Asia, and Scandinavia, which have seen artificial intelligence as a strategic opportunity to diversify their economies and position themselves as global actors in the next technological revolution. Funds such as the Saudi Public Investment Fund, Mubadala from the United Arab Emirates, Singapore’s GIC, or Norway’s Norges Bank Investment Fund have allocated tens of billions of dollars to investments in AI companies, computing infrastructures, and data centers. While this intervention has enabled accelerated sectoral growth, it has also increased vulnerability to political decisions and geoeconomic circumstances. When financing depends on national interests or international prestige objectives, investment decisions tend to prioritize visibility over profitability, creating a reputation-driven economy rather than one based on innovation.

The third pillar of this dependence comes from governments. Since 2020, the race for global leadership in artificial intelligence has acquired a political and geostrategic component. The United States, the European Union, China, Japan, and South Korea have announced public investment plans that together exceed 500 billion dollars by 2030. These programs, initially oriented toward scientific and educational development, have ended up becoming direct and indirect subsidy vehicles for the tech industry. Subsidies for data centers, tax credits for generative AI projects, and public procurement of “intelligent” software fuel the expansion of the sector without always requiring verifiable results. In many cases, national policies act as incentives for artificial demand: public markets are created to justify private growth, shifting the risks to taxpayers.

The overdependence on these three types of capital — speculative private, strategic sovereign, and institutional public — creates structural distortion. Because growth is not driven by real market demand or tangible benefits, the industry’s expansion resembles a financial bubble more than a process of technological consolidation. When capital is allocated based on expectation rather than value, innovation becomes subordinate to rhetoric. Funds demand spectacular announcements, governments seek headlines that signal leadership, and companies adapt their communication to both audiences, prioritizing media visibility over deep research. The result is an economy of technological spectacle in which funding rewards the promise before the achievement.

This dependence also has long-term consequences. If the flow of capital slows — whether due to changes in interest rates, geopolitical tensions, or a crisis of confidence — many artificial intelligence companies will lack mechanisms for self-sufficiency. The cost structure of AI models, based on increasing energy and computational consumption, makes their sustainability without external investment largely unfeasible. A contraction of capital could trigger a domino effect of closures and layoffs similar to the collapse of the tech sector after the burst of the dot-com bubble in 2001.

Beyond financial risk, this overdependence also limits scientific independence. Research projects become oriented toward commercial or geopolitical goals, and the direction of technological progress is determined by the interests of those who finance it. Priorities cease to be ethical, social, or human, and instead become economic and strategic. In this sense, excess capital can become an obstacle to real progress: innovation loses depth when short-term return takes precedence over long-term exploration.

Economic history shows that no industry can sustain itself indefinitely on the basis of subsidies or speculation. Artificial intelligence, by depending so heavily on these financing mechanisms, risks losing its autonomy and credibility. If massive investment does not translate into productivity, added value, and sustainable profits, the same capital that currently drives it will be the one that triggers its fall. Because every bubble is ultimately fed by the same thing: the belief that money will never stop flowing. Along with people’s ignorance.

3.6. The hidden energy and environmental cost: the electricity and water consumption of data centers

One of the most silenced aspects in the triumphalist narrative surrounding artificial intelligence is its enormous energy and environmental cost. While public discourse presents it as an “immaterial” technology based on algorithms and data, the physical reality that sustains its operation is monumental. Behind every generative AI model, every chatbot and every automated recommendation, there are gigantic data centers that consume massive amounts of electricity and water, relying on industrial infrastructures whose ecological footprint is growing at an alarming rate. This cost, almost always absent from media conversations, is one of the most critical and least sustainable factors in the current rise of artificial intelligence.

The training and operation of large scale AI models require colossal computational infrastructure. Every state of the art language model or computer vision system needs tens of thousands of graphics processing units working continuously for weeks or even months. According to estimates from the University of Massachusetts Amherst, training a single language model with 175 billion parameters can generate carbon dioxide emissions equivalent to those produced by five cars during their entire lifetime. OpenAI, for example, used an estimated consumption of more than 1,200 megawatt hours to train GPT 4, enough to power a city of 30,000 inhabitants for several days. As companies compete to train increasingly large and complex models, the energy cost grows exponentially.

Electricity consumption is only part of the problem. The data centers that host the servers require constant cooling systems to prevent the equipment from overheating. This process uses millions of liters of water every year. A report from the University of California, Riverside revealed that the training of GPT 3 consumed around 700,000 liters of water, and that the daily use of generative AI models by millions of users could raise that figure to levels comparable to the annual consumption of small cities. In countries where water resources are scarce, such as Spain, Saudi Arabia or certain regions of the United States, this intensive water use aggravates existing tensions between the tech industry and environmental management.

Large tech corporations, aware of their environmental impact, have attempted to mitigate their public image through greenwashing strategies. Microsoft, Google, Amazon and Meta have announced commitments to operate with 100 percent renewable energy in the coming decade. However, real data shows that although some facilities are partially powered by clean sources, total energy demand is growing faster than the capacity to compensate for it. In 2025, the International Energy Agency estimated that data centers linked to AI would consume more than 1,000 terawatt hours per year, equivalent to the combined electricity consumption of Japan and the United Kingdom. Even if part of that energy comes from renewable sources, the sheer scale of total consumption remains a structural challenge for global sustainability.

Another aspect seldom mentioned is the dependence on highly polluting materials and industrial processes. The manufacture of advanced chips, particularly those used in GPUs and TPUs, requires rare metals and enormous amounts of ultrapure water. Semiconductor factories, concentrated in countries such as Taiwan, South Korea and the United States, consume water volumes comparable to those of large urban complexes, while generating chemical waste that is difficult to process. The advancement of AI therefore carries not only a direct energy cost but also an indirect ecological cost linked to the production of the hardware that makes it possible.

The central contradiction is that artificial intelligence is presented as a tool to fight climate change, improve energy efficiency and optimize natural resources, while its own operation exacerbates the problems it claims to solve. Energy optimization or environmental management algorithms, for example, run on systems that consume more energy than they save. Likewise, AI projects for climate forecasting or sustainable agriculture depend on infrastructures that increase global emissions. This imbalance between discourse and reality is one of the clearest signs of the disconnect between technical progress and ecological sustainability.

The problem of the energy and environmental cost of artificial intelligence goes beyond technological efficiency: it reflects the economic model that drives it. The logic of unlimited growth, training ever larger models, processing more data and offering faster responses, is in direct conflict with the need to reduce global energy and water consumption. Companies justify this waste as a “necessary investment for the future”, but that future depends on finite resources that are already in crisis. If the sector does not redefine its priorities toward real efficiency and verifiable sustainability, the environmental impact of AI could become the trigger for a new global contradiction: that of a digital revolution which, in the name of intelligence, accelerates the degradation of the planet that sustains it.

The hidden energy and environmental cost of artificial intelligence shows that true intelligence does not lie in the ability of algorithms to generate text or images, but in the human capacity to recognize the limits of its own model of progress. As long as industry and governments continue to measure success in terms of scale and computational power, the ecological footprint of AI will continue to grow silently, becoming the invisible price of a revolution that promises to change everything but may be mortgaging the future of the very planet on which it depends.

3.7. The geopolitical factor: the technological war between the United States and China as a catalyst for irrational spending

The development of artificial intelligence has become not only an economic and technological race, but also a geopolitical competition that is reshaping the global balance of power. The rivalry between the United States and China for dominance in AI has escalated to levels comparable to the space race of the Cold War, although this time the battle is not fought in the skies, but in laboratories, chip factories, and data centers. This confrontation, driven by strategic ambitions and the pursuit of economic supremacy, has become one of the main forces fueling irrational spending, inflated investments, and the unrestrained acceleration of the sector.

For both countries, artificial intelligence represents far more than a technological tool: it is an asset of national sovereignty. The United States sees it as the key to maintaining economic and military hegemony in the twenty first century, while China sees it as the path to consolidating itself as the dominant power in the new global technological order. In this context, AI has become a matter of national security. Governments, companies, and investment funds operate under the logic that if the other side invests, they must invest even more. This mirror effect has created an unprecedented spiral of spending, where the priority is not profitability or sustainability, but speed.

In the case of the United States, momentum comes from an alliance between the private sector and the military industrial complex. Large corporations such as Microsoft, Google, Amazon, Meta, Nvidia, OpenAI, and Palantir, among others, concentrate most of the global AI infrastructure and are direct providers of services to the government and the Department of Defense. Programs such as the Joint Artificial Intelligence Center (JAIC) or DARPA’s AI Next Campaign reflect the goal of integrating AI at all levels of defense, logistics, and strategic analysis. This public private collaboration has channeled billions of dollars into technology companies, many of which operate simultaneously in the civilian and military markets. In 2025, the United States budget for AI and advanced technology projects exceeded 80 billion dollars, a figure that increases year after year.

China, for its part, has responded with an equally ambitious plan. Since the announcement of its National Artificial Intelligence Development Plan in 2017, the Chinese government has invested colossal sums in research, education, industrial automation, and digital surveillance. Chinese companies such as Baidu, Tencent, Alibaba, SenseTime, and Huawei act as technological arms of the state, receiving direct financial support and preferential access to infrastructure and data. Beijing’s official goal is clear: to become the world leader in artificial intelligence by 2030. According to data from China’s Ministry of Science and Technology, accumulated spending on AI since 2018 exceeds 250 billion dollars, a figure that includes both public and private investment.

This competition has extended into the field of semiconductors, energy, and data infrastructure, forcing a global reorganization of supply chains. The restrictions imposed by Washington on the export of advanced chips and lithography equipment to China, especially after the 2022 CHIPS and Science Act, have triggered a parallel race toward technological self sufficiency. China has responded with massive investments in the development of its own semiconductor industry and with the creation of strategic alliances with countries in Asia and the Middle East to secure access to critical materials and energy resources. This confrontation not only increases production and development costs, but also multiplies infrastructure duplication: each power seeks to build its own closed technological ecosystem, independent from its rival.

The economic impact of this technological war is immense. According to estimates from the World Bank, the combined spending of the United States and China on research and development of artificial intelligence will exceed 700 billion dollars between 2024 and 2030. This figure, while driving innovation, also represents a form of structural inflation: capital is not allocated to generate immediate returns, but to maintain geostrategic positions. Investment is made not to gain productivity, but to avoid losing power. In this context, efficiency ceases to be a goal, and investment becomes an end in itself.

The most direct consequence of this rivalry is the artificial acceleration of the global AI market. Startups, universities, and research centers benefit from the flow of public and private funds, but they are also trapped in a logic of geopolitical competition that prioritizes speed over ethics and military impact over sustainability. Technological development is increasingly oriented toward applications in surveillance, information control, automated defense, and cyberwarfare, pushing scientific and social research into the background. At the same time, other countries such as India, Japan, South Korea, Germany, France, or Saudi Arabia are trying to position themselves in this global game, multiplying projects, consortia, and investments in an attempt not to be sidelined in the new technological landscape.

The result is a global system of excessive spending sustained by the geopolitics of fear. Every advance by the rival is interpreted as an existential threat, and every domestic investment as a measure of survival. This collective psychological mechanism is what fuels the most durable bubbles: the belief that not investing means losing, even when investing does not guarantee winning. In the case of artificial intelligence, this logic has turned innovation into a kind of economic and digital arms race, where economic rationality has been replaced by the instinct for power.

The technological war between the United States and China has driven not only the progress of artificial intelligence, but also its inflation. Strategic competition has created an economy of excess in which resources are allocated more for political reasons than for technical or social needs. Like every rivalry between great powers, this race may sustain the momentum of the bubble for a while, but it cannot prevent its eventual outcome. When returns fail to match spending and innovation no longer justifies investment, the geopolitical confrontation that today fuels the age of AI could become the catalyst for its most severe correction.

4. The role of hype and sensationalism

The rise of artificial intelligence cannot be understood without examining the role that hype and sensationalism have played in its global expansion. Beyond scientific achievements and real technological developments, the public perception of AI has been shaped by a media and corporate machinery that has amplified its capabilities, distorted its limitations, and turned every announcement into a historic event. In this narrative, artificial intelligence is not presented as an evolving tool but as an almost supernatural force destined to redefine the world.

Hype, driven by economic, political, and marketing interests, acts as an emotional multiplier: it turns curiosity into enthusiasm, enthusiasm into investment, and investment into conviction. Media outlets, consulting firms, opinion leaders, and the self-proclaimed “experts” on social networks have contributed to creating an environment where exaggeration is perceived as innovation and doubt as resistance to progress. As a result, the discourse on artificial intelligence has become filled with impossible promises, apocalyptic predictions, and utopian visions, generating a blend of fascination and fear that keeps the constant attention of both the public and the markets.

Analyzing the role of hype and sensationalism is essential to understanding how artificial intelligence has evolved from a specialized scientific field into a global cultural phenomenon. It is precisely in this tension between technical reality and media fiction where much of the current bubble is sustained. The following section examines how the media, consulting firms, and self-appointed experts have inflated expectations, fueling a dynamic in which the perception of progress far exceeds the truth of the facts.

4.1. How the media, consulting firms, and “LinkedIn experts” inflate expectations

The media, major consulting firms, and the so-called “LinkedIn experts” play a central role in creating and perpetuating the hype surrounding artificial intelligence. They are largely responsible for turning legitimate technical advances into media phenomena and converting scientific complexity into mass-market slogans. Their influence, amplified by social networks and digital information channels, has produced a narrative in which every new model, algorithm, or application is presented as a historic leap comparable to the invention of electricity or the discovery of DNA. This discourse, repeated to the point of saturation, has inflated expectations to a level no technology could meet without resorting to exaggeration or fantasy.

Traditional media act as the first link in this amplification chain. In their constant search for striking headlines, they simplify scientific advances until they become emotional, personalized, and spectacular stories. A system capable of generating images or text, despite being limited and dependent on preexisting data, is described as “a machine that thinks”, “an artificial brain”, or “an intelligence beyond the human one”. The boundaries between technical description and metaphor vanish, and the language of science is replaced with the language of myth. The result is a distorted public perception in which AI is seen not as a set of algorithmic tools, but as an emerging form of life endowed with intentionality and consciousness.

Large consulting firms, for their part, have found in this narrative an inexhaustible source of business. Organizations such as McKinsey, Deloitte, PwC, or Accenture constantly publish reports, forecasts, and impact studies projecting multi-trillion-dollar figures on the potential economic value of artificial intelligence. These reports claim that AI will add between 13 and 15 trillion dollars to global GDP in the next decade, that it will transform all productive sectors, and that it will generate a “new economy of intelligence”. Rarely do these documents include verifiable methodologies or critical considerations regarding sustainability or inequality, yet they serve as rhetorical ammunition to justify massive investments and corporate restructurings. Their function is not so much to inform as to persuade; they sell certainties in a context where uncertainty is abundant.

To this media and corporate machinery we must add a new actor of enormous influence: the “LinkedIn expert”. This figure, a product of contemporary digital culture, has turned public commentary on AI into a form of personal marketing. Professionals without deep scientific or technical training publish daily superficial analyses, motivational phrases, and grandiose conclusions about the future of humanity, accompanied by hashtags and generic statistics. Their goal is not to disseminate knowledge but to capture attention. Virality has replaced rigor, and the number of reactions or followers has become a measure of credibility. This phenomenon has democratized the discourse but also degraded it: complexity dissolves into empty statements such as “AI will not replace humans, but those who don’t use it”, which circulate millions of times without providing any real content.

The combination of media, consulting firms, and social-media experts has created a perfectly self-reinforcing ecosystem. Consulting firms publish optimistic reports that the media cite as unquestionable truths; the media produce sensationalist headlines that influencers repeat on their profiles; and these influencers amplify the message among millions of users who, in turn, share it as a personal revelation. Within days, a prediction without empirical basis can become a global consensus. This cycle of constant amplification not only distorts public perception but also shapes political and business decisions, driving rushed investments, ineffective regulations, and corporate strategies based more on the fear of falling behind than on a rational assessment of risks and benefits.

A revealing example of this phenomenon can be seen in the media coverage of generative artificial intelligence since 2023. Headlines such as “ChatGPT will change education forever”, “AI will write all future laws”, or “Lawyers will be replaced by algorithms” dominated global conversation for months, even though none of these claims were supported by verifiable data or empirical evidence. Yet the psychological impact was immediate: millions of people began to believe their professions were at risk, and thousands of companies launched improvised AI projects so as not to appear obsolete. What began as a media narrative ended up altering economic reality.

This discursive inflation has profound costs. By turning AI into a permanent spectacle, the media and consulting firms have shifted attention away from the truly important topics —ethics, sustainability, technological dependency, education, and employment— and toward an emotional space dominated by promise and fear. Society has been persuaded that we stand on the brink of an imminent revolution when, in many cases, what actually exists is gradual evolution with partial results and significant technical limitations. The paradox is that by inflating expectations to impossible levels, the very actors promoting today’s euphoria will contribute tomorrow to collective disappointment when reality fails to fulfill what was promised.

Hype does not arise from ignorance but from interest. The media need attention, consulting firms need clients, and self-proclaimed experts need visibility. Together, they have turned artificial intelligence into the most profitable narrative of the twenty-first century. And as with every previous bubble, that narrative holds as long as faith in the promise remains stronger than evidence of its limits. When that faith begins to erode, the discourse of miracles will give way to the discourse of disappointment, and the noise of hype will be replaced by the silence of correction.

4.2. Mass psychology and the contagion effect: FOMO, promises of instant success, the idea that if you are not using AI you are already obsolete

The phenomenon of the artificial intelligence bubble cannot be explained solely through economics or technology; it is, above all, a collective psychological phenomenon. At the center of this dynamic lies mass psychology and its ability to amplify perceptions, spread emotions, and turn a trend into a global conviction. AI has become a symbol of success, power, and modernity, and around it an emotional narrative has emerged that activates the same mechanisms seen in other major historical bubbles: the fear of being left behind, the promise of immediate success, and the illusion of inevitability. The result is a collective mindset in which the decision to adopt or invest in artificial intelligence does not always stem from rational analysis, but from a deeply rooted social impulse: FOMO, or fear of missing out.

Let me clarify what FOMO means. FOMO is the invisible engine driving much of the investment and accelerated adoption of AI. Governments, companies, and professionals feel pressured by the idea that anyone who does not integrate artificial intelligence into their processes is doomed to disappear. This narrative, the idea that if you are not using AI you are dead, has become a mantra repeated endlessly by consultancies, influencers, and specialized media. As a result, technological adoption becomes a question of symbolic survival rather than one of real necessity or proven efficiency. Artificial intelligence is not implemented because it is understood, but because people fear the consequences of not doing so. This fear, more than hope in innovation, is what is accelerating the uncontrolled expansion of the market.

The psychological contagion effect functions like a chain reaction. When a company announces that it is implementing AI, its competitors feel obliged to do the same, even when no clear strategy or demonstrable return exists. Similarly, governments launch national artificial intelligence plans when they see other countries doing so, universities create master's degrees and specialized programs to avoid appearing outdated, and professionals fill their profiles with certifications and references to AI to show that they are keeping up. The phenomenon is not based on knowledge, but on perception. And perception, in an attention economy, is worth more than truth.

In the corporate world, this collective psychology has generated a culture of rushed adoption. Companies invest in AI systems without defined objectives, purchase expensive licenses they do not fully use, or integrate chatbots and digital assistants without evaluating their real impact on productivity. Instead of optimizing, many organizations end up duplicating processes and creating technological inefficiencies. Yet the fear of being seen as lagging behind weighs more than evidence and results. In some cases, executives openly acknowledge that their investment in AI responds more to market pressure than to strategic planning.

At the individual level, FOMO manifests itself as generalized professional anxiety. Workers in all sectors feel that they must learn, master, or incorporate AI tools in order not to be replaced. Courses, certifications, and training programs proliferate as promises of immediate success. Educational platforms, training companies, and digital coaches exploit this fear by offering quick and miraculous solutions for not falling behind. The promise of instant success, such as claims that using AI will double your productivity or that mastering ChatGPT will earn you thousands of euros per month, replicates the same psychological logic found in pyramid schemes or investment fads. What matters is not transformation, but the illusion of belonging to the cutting edge.

Mass psychology thus turns artificial intelligence into a social phenomenon before a technological one. The desire to belong to the future generates an imitation dynamic that goes beyond reason. Human beings, as in any phenomenon of emotional contagion, move in groups: external validation replaces internal evaluation. If everyone is doing it, it must be right. This principle of herd behavior explains why entire sectors rush to adopt incomplete technologies, why the media amplify each new model as if it were a revolution, and why companies spend millions on systems they do not understand. Emotional contagion, more than rational planning, becomes the dominant force.

The narrative that if you are not using AI you are dead synthesizes this collective mindset. It is not a technical statement but a cultural construct designed to generate urgency. Its effectiveness lies in appealing simultaneously to fear and desire: fear of irrelevance and desire for superiority. Like all narratives driven by social pressure, this dichotomy produces both enthusiasm and anxiety, and both emotions sustain the same market. The individual or company that does not adopt AI is not perceived as prudent but as outdated; and in an environment where the image of modernity is worth more than evidence of results, that perception is enough to justify the expense.

Historically, every period of technological euphoria has reproduced this same dynamic. During the dot-com bubble, investors feared missing the next Microsoft; during the 2008 crisis, banks feared being left out of the new financial instruments; today, companies fear being left out of the intelligent future. In all cases, the fear of not participating was more powerful than rationality when evaluating risks. And in all cases, the outcome was the same: when massive enthusiasm collided with reality, emotional contagion turned into financial contagion.

Mass psychology and FOMO are the invisible pillars of the artificial intelligence bubble. Social pressure to stay up to date and the fear of exclusion have replaced reflection with imitation. And as long as AI adoption continues to be guided by collective emotions rather than conscious strategies, the market will continue growing on the basis of anxiety. Because in this new economy of intelligence, the most profitable product is not the algorithm, but the fear of being left behind.

4.3. The role of tech companies and their communication departments in creating “miracles” that do not exist

Large technology companies have played a decisive role in shaping the artificial intelligence bubble by becoming not only developers of technology but also manufacturers of narratives. Their communication, marketing, and public relations departments have learned to operate with the precision of a global propaganda machine, capable of transforming any incremental advance into a historic event. The result is a media ecosystem in which the boundaries between innovation, fiction, and advertising blur, and where the perception of progress matters more than progress itself. These corporations do not simply sell products or services; they sell a vision of the future, carefully designed to inspire fascination, dependency, and above all, trust.

The creation of technological “miracles” begins with the manipulation of language. Terms such as “intelligence”, “reasoning”, “creativity”, or “understanding” are deliberately used to describe algorithmic processes that possess none of those qualities in a human sense. A predictive language model that generates statistically coherent text is presented as a “system that thinks”, an image recognition program is described as “able to see”, and a digital assistant that follows preprogrammed scripts is advertised as “able to understand”. This anthropomorphic language, repeated systematically in press releases, conferences, and advertisements, creates the illusion that artificial intelligence has reached levels of consciousness or intention that, in reality, do not exist.

The communication departments of major tech companies operate with the same logic as political or religious campaigns: they do not aim to inform but to mobilize emotions. Every presentation of a new AI model or product is staged as a revelation, accompanied by visionary speeches, visual effects, and phrases carefully designed to remain in the collective memory. “We are redefining the way the world thinks”, “intelligence is no longer exclusively human”, or “we have taken the first step toward AGI” are expressions that appeal directly to the imagery of the transcendental. The strategy is clear: replace rational evaluation of technical progress with the feeling of witnessing a historic event.

This communication strategy has turned the conferences of major tech companies into global spectacles. Product launch events for systems like ChatGPT, Gemini, Copilot, or Claude are presented with a theatricality reminiscent of a movie premiere or a cult ritual. Carefully selected demonstrations, promises about future capabilities, and declarations about “imminent revolutions” are combined. However, most of these demonstrations are controlled down to the last detail, based on pre-trained examples, and do not show the real limitations of the systems. What is displayed is not the technology in its authentic state, but its idealized version, designed to reinforce the narrative of the miracle.

The purpose of these campaigns is not only to attract users but to maintain media attention and investor confidence. In a market where stock value depends more on expectations than on results, communication becomes the central instrument of speculation. Every announcement of a new model, every mention of “general AI”, or every viral demonstration generates an immediate increase in market capitalization, investment, and media coverage. This cycle of announcements and reactions keeps the sector’s value inflated, even when the real profitability of the products remains uncertain. The economy of faith replaces the economy of results.

At the same time, companies use the strategy of opacity to reinforce the aura of mystery surrounding their advances. The lack of transparency regarding training data, evaluation methods, or technical limitations is justified as “intellectual protection”, but in reality it serves to maintain the myth. By not revealing the magnitude of biases, energy costs, or model fragility, corporations ensure that public perception is based more on fascination than on understanding. This opacity prevents scientific and ethical scrutiny but strengthens the narrative of absolute power: the idea that the technology is so advanced that it is incomprehensible to ordinary people.

The case of OpenAI and the launch of ChatGPT in 2022 is emblematic. The model was presented as a qualitative leap toward general intelligence, but the technical reality was far more modest: a statistical system based on text prediction trained on large volumes of data. However, the media impact was such that millions of people perceived it as the beginning of a new era. Microsoft, Google, and other companies responded immediately with their own versions, fueling a cascade of announcements. Each presentation generated new headlines, new investors, and new expectations, regardless of whether the advances were truly significant or simply incremental evolutions.

The pattern is always the same: create the feeling of standing on the verge of a transcendental discovery and at the same time maintain the promise that the best is yet to come. This balance between the present miracle and the future hope allows companies to sustain a permanent narrative of progress. Every technical limitation is reinterpreted as a challenge that will be overcome “in the next version”, keeping attention and capital flowing without interruption. In this way, communication departments have managed to turn scientific uncertainty into commercial fuel.

The problem is that this systematic construction of miracles has profound consequences. By inflating expectations beyond what is reasonable, tech companies distort public perception and erode long-term trust. When reality does not match the promises, disillusionment turns into cynicism, and enthusiasm into distrust. But in the meantime, the benefits have already materialized: shares have risen, contracts have been signed, and media attention has been secured.

Believing in progress, in technological salvation, in the idea that humanity is heading toward an era of algorithmic perfection, and as with any belief, what sustains the system is not truth but faith. The communication departments of major tech companies have understood this better than anyone: they do not need to create artificial intelligence to dominate the world; they only need to create the illusion that they already have. Creating dependency.

4.4. Concrete examples: announcements of imminent AGI, robots with emotions, models that “think” or “feel”

Among the many mechanisms that fuel the artificial intelligence bubble, few are as obvious as the constant production of spectacular announcements about achievements that do not actually exist. The tech industry, driven by the need to maintain media attention and investor confidence, has turned into a routine the presentation of “historic breakthroughs” that promise the imminent arrival of Artificial General Intelligence (AGI), the creation of robots capable of feeling human emotions, or the development of models that supposedly “think” and “reason”. These narratives, carefully crafted, are not the result of a communication error but a deliberate strategy to sustain the storyline of the technological miracle and perpetuate the cycle of hype.

Announcements about the proximity of AGI, the hypothetical artificial intelligence capable of autonomous reasoning, learning, and adaptation at a human level, are perhaps the clearest example. Since 2022, executives from companies like OpenAI, Google DeepMind, and Anthropic have made public statements claiming that AGI could be “a few years away”. Sam Altman, CEO of OpenAI, has repeatedly stated that his company “was born to create AGI and ensure it benefits all of humanity”, while Demis Hassabis of DeepMind has declared that his team “is closer than ever to achieving human-level intelligence”. These claims, devoid of any clear scientific definition of what AGI actually means, operate in the realm of faith and rhetoric. No current system possesses the minimum characteristics that would define general intelligence: self-awareness, abstract reasoning, or causal understanding of the world. However, the mere act of announcing its imminence is enough to keep alive the perception of an uncontrollable revolution underway.

These announcements have immediate effects on the market and public opinion. After every declaration of “progress toward AGI”, the stock valuations of the companies involved rise, investors double down on their bets, and the media amplify their coverage. In this way, AGI, still a philosophical and theoretical concept, becomes a financial product. Speculation no longer revolves around real technology but around the expectation of a miracle. It follows the same psychological pattern that characterized past bubbles: the replacement of evidence with promise.

The same phenomenon appears in announcements of robots “with emotions”. Robotics companies such as Hanson Robotics, Boston Dynamics, and Engineered Arts have repeatedly showcased humanoids designed to mimic human facial expressions and tones of voice, promoting them as “empathetic machines” or “emotional entities”. Sophia, the famous robot from Hanson Robotics, even received symbolic citizenship in Saudi Arabia in 2017, in a marketing stunt that made headlines around the world. However, Sophia does not possess emotions, understanding, or consciousness; its responses are pre-programmed scripts combined with basic natural language processing techniques. Its “expressiveness” is the result of motors and sensors designed to reproduce human micro-gestures. Nevertheless, the public, conditioned by media discourse and anthropomorphic appearance, perceives these demonstrations as evidence of a qualitative leap toward artificial sentience.

The case of generative language models such as ChatGPT, Gemini, or Claude follows the same logic. With each new version, press releases describe them as “models that think”, “reason”, or “understand the world”, when in reality they are statistical systems that predict the next word in a sequence based on massive volumes of data. These models possess neither consciousness nor understanding; they operate through mathematical correlations, not semantic comprehension. However, the fluid and convincing way they generate text or imitate human emotions creates a cognitive illusion that corporate marketing deliberately exploits. It plays with the confusion between behavior and consciousness, between coherent response and thought. Thus, an illusion of intelligence is created where there is only probabilistic calculation.

To this strategy we must add the constant announcements of “revolutions” that never materialize. AI companies promise advancements in fields such as medicine, education, or justice that, in practice, never move beyond the experimental stage. They announce systems that diagnose diseases better than doctors, platforms that will eradicate hiring biases, or programs that will write “fairer and more rational” laws. However, in most cases, the results published in academic articles or technical reports are far more modest: high error rates, persistent biases, or dependence on incomplete data. The headlines survive; the technical corrections go unnoticed.

The use of controlled demonstrations and carefully edited videos is another key component in the creation of artificial miracles. Companies showcase prototypes performing exceptional tasks under perfectly designed conditions: robots that dance, systems that generate “original” music, virtual assistants that respond with emotional tone, or autonomous vehicles that drive flawlessly. What remains hidden are the limits: the repeated errors in uncontrolled contexts, the dependence on specific datasets, or the failures in real-world scenarios. The media spectacle, rather than scientific experimentation, has become the primary tool of legitimization.

These technological “miracles” serve both economic and symbolic functions. In the short term, they attract capital, generate headlines, and reinforce the companies’ position as leaders of the future. In the long term, they build a cultural narrative in which artificial intelligence is portrayed as an inevitable and autonomous force capable of surpassing the human condition. This narrative does not need to be true to be effective; its power lies in the emotion it evokes. The public does not evaluate the technical accuracy of a model but the sense of wonder it produces.

The consequences of this game of illusions are twofold. On the one hand, it creates an environment of impossible expectations in which any technological slowdown is perceived as a failure or betrayal. On the other hand, it undermines the credibility of science and engineering when these promises go unfulfilled. Ultimately, the spectacle of imminent AGI, emotional robots, and models that “think” does not aim to inform or advance knowledge, but to keep global attention alive. Tech companies have understood that in the information age, perception is the most valuable product. And as long as humanity continues to confuse emotion with intelligence and marketing with progress, miracles will keep selling faster than truth.

4.5. The contrast with technical reality: limitations of LLMs, the lack of semantic understanding, and the dependence on data

The contrast between the media narrative surrounding artificial intelligence and its technical reality is immense. While tech companies and the media present large language models as almost conscious systems capable of reasoning, creating, and understanding, the scientific truth is far more limited and prosaic. LLMs, which constitute the foundation of the current generative AI revolution, are not thinking entities nor systems of deep understanding. They are statistical prediction algorithms trained on enormous quantities of text. Their apparent intelligence does not come from any real understanding of the world, but from the ability to calculate the most probable word that should follow another within a linguistic sequence.

LLMs such as GPT, Gemini, Claude, or LLaMA operate under an architecture known as a transformer, which makes it possible to process large volumes of textual information and detect complex patterns of relationships between words. These models possess neither consciousness nor intention, nor any notion of meaning. They do not know what they are saying; they simply generate coherent sentences based on learned statistical correlations. Their ability to imitate human language is based on the size of the dataset and the computational power employed, not on any understanding of the content. In technical terms, what they achieve is a sophisticated simulation of language, not the reproduction of intelligence.

The absence of semantic understanding is one of the most evident limits. The models can describe an emotion but cannot feel it; they can explain a philosophical concept but cannot understand it; they can generate computer code without knowing what it is for or what consequences it may have. Their knowledge is purely syntactic, a repetition of learned structures and associations. This becomes evident in errors known as hallucinations, responses that seem convincing but are completely false or fabricated. A model may cite nonexistent studies, invent biographies, or mix facts with fiction without being able to distinguish the difference. The reason is simple: it has no representation of the world, only of the texts on which it was trained.

Moreover, LLMs depend entirely on the data they consume, which implies a series of structural limitations. First, they reproduce the biases, errors, and inequalities present in their sources. If the dataset contains racial, gender, or cultural biases, the model will replicate them, even when attempting to avoid them. Second, their knowledge is static: what they learn becomes frozen at the time of training. Unlike humans, who update their understanding based on experience, language models must be retrained from scratch to incorporate new information, a costly and energy-intensive process. Consequently, their knowledge is always out of date with respect to reality, making them imitators of the past rather than interpreters of the present.

Another fundamental limitation lies in the superficial nature of their learning. LLMs do not understand contexts, intentions, or the implications of words. A classic example is their difficulty in reasoning with common sense or handling contradictions. When presented with a paradox or an ambiguous question, they tend to produce responses that sound plausible but lack internal logic. Likewise, they are incapable of formulating coherent value judgments, establishing causal relationships, or anticipating consequences beyond the linguistic patterns available to them. This type of symbolic reasoning, inherent to human thought, remains beyond their reach.

The problem becomes more severe when society, and especially institutions, begin to make decisions based on the outputs of these models. In environments such as education, medicine, or justice, where interpretation and ethics are essential, dependence on systems without real understanding poses serious risks. An AI model can draft legal reports or medical diagnoses that appear precise but are based on faulty correlations or incomplete data. In the hands of users who do not understand its internal functioning, the illusion of competence can lead to serious errors. Blind trust in these systems is not a sign of progress but a renunciation of critical thinking.

The technical limitations are also reflected in the dependence on hardware and computational resources. Training the most advanced models requires enormous infrastructures, thousands of GPUs, and colossal amounts of energy. This restricts the development of advanced AI to a handful of corporations with access to capital and infrastructure, further consolidating the concentration of technological power. Far from democratizing knowledge, LLMs are reinforcing a new form of digital inequality: knowledge mediated by the corporations that control the data and algorithms.

Despite these known and documented limits, corporate and media discourse continues to promote the idea that these models are close to achieving human intelligence. However, the real challenge lies not in the scale of the data nor in processor speed, but in the nature of learning and understanding. No increase in parameters or model size can substitute for the absence of meaning. As long as AI cannot build conceptual representations of the world, an unbridgeable distance will remain between imitation and understanding.

The contrast between the myth and the technical reality reveals an uncomfortable truth: the artificial intelligence that dominates the market today is, in essence, a machine of predictions, not of knowledge. Its ability to generate text, images, or music with an appearance of understanding is nothing more than a statistical reflection of the human data that feeds it. What is perceived as thought is only probability; what is interpreted as creativity is combination; and what is sold as consciousness is calculation. The danger does not lie in what the models can do, but in what humanity is willing to believe they do.

4.6. How apocalyptic and utopian narratives feed each other (fear and hope sell equally)

The discourse around artificial intelligence is trapped in a paradox that sustains and perpetuates it: the coexistence of fear and hope as forces of persuasion. Both extremes, the apocalyptic and the utopian, feed each other in a continuous cycle that keeps AI at the center of the global conversation, generating both fascination and anxiety. This duality is not accidental but the result of an effective communicative strategy that exploits the most universal human emotions: the fear of losing control and the desire to transcend one’s limits. In the contemporary imagination, AI is simultaneously the executioner and the savior, the threat that could destroy civilization and the promise that could elevate it to perfection.

The utopian narrative presents artificial intelligence as the culmination of human progress. In this vision, algorithms will solve hunger, cure diseases, eliminate poverty, and optimize every aspect of existence. Machines endowed with superior intelligence will bring a new era of abundance and unlimited knowledge. This storyline, driven by tech companies, investors, and future-focused gurus, relies on a quasi-religious narrative: the idea that technology will redeem us from our own mistakes. AI thus becomes a new secular divinity, a perfect system that, unlike humans, does not tire, does not corrupt, and does not make mistakes. The promise of a civilization governed by algorithmic reason is sold as the next evolutionary leap.

On the opposite extreme lies the apocalyptic narrative, fueled by the same fascination. In this storyline, artificial intelligence represents an existential threat: an autonomous force that will end up dominating or exterminating humanity. It is associated with scenarios of massive job loss, political manipulation, collapse of privacy, or even total annihilation. Figures such as Elon Musk or the late Stephen Hawking publicly warned about the dangers of a “superintelligence” that might escape human control. Films, novels, and series reinforce this image with visions of dystopian futures where machines rule, humans serve, and consciousness becomes a luxury of the past.

Although they appear to be opposites, both narratives serve the same function: to keep attention and investment flowing. Hope attracts optimists; fear mobilizes skeptics. Tech companies use utopia to sell promises and apocalypse to justify the urgency to act. Thus, the same actors who proclaim that AI will save humanity also warn that, without their guidance, it could destroy it. This double discourse creates an emotional dependency: the public oscillates between admiration and panic, and in both cases continues to consume information, products, and solutions that feed the cycle.

The media have perfected this dialectic. Every technological breakthrough is presented with a dual headline: “AI could cure cancer” appears alongside “AI could eliminate millions of jobs.” The goal of the narrative is not to inform but to provoke. Fear and hope alternate as complementary stimuli in a strategy of constant attention. Market analysts and consulting firms reinforce the same pattern: they publish studies that celebrate the economic benefits of AI while simultaneously warning about its destructive potential. This ambiguity keeps the public trapped in an exciting uncertainty in which technology is seen as inevitable but also unpredictable.

The apocalyptic and utopian narratives share a common root: both are built on exaggeration and simplification. At their core, they are two sides of the same storyline that strips humans of agency. In the utopian narrative, machines liberate us; in the apocalyptic one, they replace us. In neither case does the human being retain responsibility or decision-making power. This technocentric worldview turns AI into the protagonist of history and humanity into its spectator. That is precisely the symbolic function of the bubble: to create a myth powerful enough to justify massive investment, collective fascination, and the suspension of critical thinking.

Both narratives also function as mechanisms of legitimization for different interests. The utopian account fuels corporate marketing and financial speculation, while the apocalyptic one legitimizes political intervention, regulation, and the concentration of power. Companies appeal to hope to attract investment and to fear to claim authority. In this way, artificial intelligence becomes a total narrative: a storyline that encompasses both salvation and disaster and that adapts according to the narrator’s convenience.

The psychological effect of this duality is profound. Society oscillates between euphoria and anxiety, between irrational enthusiasm and technological fatalism. This emotional pendulum generates an informational dependency that prevents balance. Those who fear the future of AI seek news that confirms their fear; those who idealize it seek evidence of its greatness. In both cases, emotion prevails over reason. And while the public debates between utopia and apocalypse, the industry continues to expand without accountability, protected by the attention that both extremes guarantee.

The narrative of fear and that of hope are not opposed: they need each other. Together they form a perfect narrative ecosystem in which every scenario is profitable. If AI saves the world, the narrative will have been prophetic; if it destroys it, it will have been visionary. In either case, the story will have been successfully sold. Because in the economy of the twenty-first century, truth matters less than the emotion it generates. And in the market of artificial intelligence, fear and hope are the two most lucrative products.

5. Warning signs that are already present

Every bubble, before it bursts, emits signals that can be ignored, minimized, or interpreted as simple market fluctuations. The current artificial intelligence bubble is no exception. Although the dominant narrative continues to revolve around enthusiasm and promises, indicators of imbalance are beginning to multiply in plain sight. Massive layoffs in startups, reductions in investment, failed projects, market saturation, drops in valuations, and an evident exhaustion of both technical and human resources reveal a landscape reminiscent of the final stages of other major technological manias.

These signals are neither isolated nor temporary: they are structural symptoms of a model that is expanding faster than it can sustain itself. The growth rate of the sector has far exceeded its real capacity to generate value, and the gap between expectation and outcome is becoming increasingly visible. At the same time, governments, institutions, and consumers are showing signs of fatigue toward a narrative that promised an immediate revolution but has in many cases delivered more noise than transformation.

Analyzing these signs does not mean announcing a collapse, but rather recognizing the maturity of a cycle approaching its inflection point. Identifying them clearly is essential to understanding that the AI bubble is not a future threat: it is already unfolding, silently but consistently, behind the shine of promises and headlines.

5.1. Massive layoffs in AI startups

One of the clearest signs that the artificial intelligence bubble is beginning to tighten is reflected in the massive layoffs that have shaken the technology sector over the past two years, directly affecting both large corporations and recently founded startups. What until recently was described as an infinitely expanding market, with thousands of emerging companies competing to reinvent the future, now shows a landscape of contraction, restructuring, and forced adjustment. The promise of unlimited growth, driven by a wave of venture capital and inflated expectations, has begun to collide with the reality of costs, lack of profitability, and market saturation.

Since mid 2023, dozens of artificial intelligence startups, especially those focused on generative applications, business automation, and data analysis, have reduced their workforce or shut down operations. According to data from Layoffs.fyi, more than 350 technology companies directly or indirectly linked to the AI sector carried out layoffs between 2023 and 2025, affecting more than 100,000 workers worldwide. What is most significant is that many of these layoffs are not due to abrupt drops in revenue but to the absence of sustainable business models and the exhaustion of venture capital. The investment flow that seemed inexhaustible between 2021 and 2023 has begun to slow, forcing startups to confront the reality of operating without a continuous source of external funding.

In Silicon Valley, where generative AI reached a peak of euphoria in 2023 following the launch of ChatGPT, dozens of companies created in the heat of initial enthusiasm began to shut down or reduce operations in 2024. Startups that promised to revolutionize sectors such as education, marketing, or programming through generative AI, some with valuations exceeding one billion dollars, have been forced to lay off large portions of their teams due to the lack of stable income. The most common downfall does not occur because of a lack of innovation, but because of the absence of clients willing to pay for services that, in many cases, can be replicated by free or cheaper tools from large corporations.

Europe and Asia have followed the same trend. In London, Berlin, and Paris, several emerging companies dedicated to process automation and the development of virtual assistants announced workforce reductions of more than 40 percent during the first half of 2025. In China, where the government had fueled a massive wave of AI startups through state support, market consolidation has triggered a quiet purge: thousands of small companies have been absorbed or dissolved due to their inability to compete with national giants Baidu, Tencent, and Alibaba. In India, where AI was presented as the new digital employment revolution, more than 200 startups have closed since late 2024, many of them unable to scale beyond initial marketing.

Even the major tech corporations, traditionally perceived as immune to volatility, have begun adjusting their teams. Google, Meta, Amazon, and Microsoft, despite record profits, have reduced thousands of jobs in AI research and development departments. In 2024, Microsoft dismissed part of its ethics and responsible AI team; Google cut staff in DeepMind after restructuring its research divisions, and Meta dissolved parallel projects related to AI applied to the metaverse. Although officially justified as strategic realignments, these moves reflect a transition from initial enthusiasm to a phase of cost and priority rationalization.

The underlying cause of these layoffs is structural: the growth model that fueled the explosion of generative AI was not sustainable. During the boom of 2022 and 2023, the promise of automating everything attracted massive capital and triggered a race to launch products at a pace faster than reasoning. However, most startups failed to differentiate themselves technologically or establish recurring revenue streams. Their dependence on APIs from major providers such as OpenAI, Anthropic, or Google turned many of them into mere intermediaries, unable to control operating costs or provide real added value. When the price of accessing the models increased or investors began demanding concrete results, companies without solid financial fundamentals collapsed.

The phenomenon also has a psychological dimension. During the boom, working in artificial intelligence became a symbol of prestige and a guarantee of a promising future. Engineers, data scientists, and machine learning specialists were hired at unprecedented rates, with salaries doubling those of other tech sectors. Today, many of these professionals face a saturated and increasingly competitive job market where opportunities are shrinking and projects are disappearing. This employment hangover reflects a profound shift: the talent bubble, just like the financial one, is beginning to adjust.

Massive layoffs are not only evidence of a temporary crisis but the exhaustion of a cycle of euphoria. They are the first visible consequence of an inevitable correction: the shift from growth based on expectations to growth based on results. What we are seeing is not the end of artificial intelligence, but the end of the illusion that simply saying those two letters was enough to generate value. In any bubble, the first sign of fragility is not the fall of stock prices, but the loss of jobs. And in the case of AI, that signal is already lit.

5.2. Decline in value of overvalued companies

The overvaluation of artificial intelligence companies has been one of the most powerful drivers of the current tech bubble, and its gradual correction is one of the clearest signs that the speculative cycle is entering its adjustment phase. Over the past years, the expectation that artificial intelligence would transform every sector of the economy led investors, venture capital funds, and stock markets to inflate the valuations of hundreds of companies far beyond their real ability to generate profits. Today, the gap between promised value and tangible value is beginning to close, and it is doing so with inevitable consequences: stock market crashes, massive restructurings, and a widespread loss of confidence in the narrative of unlimited growth that defined the most intense years of the AI boom.

At the height of the euphoria, between late 2022 and mid-2024, the artificial intelligence market reached historic numbers. OpenAI, Anthropic, Cohere, and other generative AI companies multiplied their valuations within months, attracting multibillion-dollar funding rounds at a speed reminiscent of the excesses of the dot-com bubble. The case of OpenAI is emblematic: its valuation rose from 14 billion dollars in 2021 to more than 90 billion in 2024, even though its net profitability remained negative and its business model depended almost entirely on a single corporate partner, Microsoft. This type of growth did not reflect financial results but speculative expectations built on a narrative: that generative artificial intelligence would redefine the world as soon as it could be fully integrated into everyday life.

However, the market began to show signs of fatigue in late 2024. Financial reports from several companies revealed that revenue from subscriptions and AI model licenses was far below projections. At the same time, infrastructure costs, particularly those related to energy consumption and GPU usage, were far exceeding profit margins. The first consequences appeared in private valuations, which began to adjust quietly. Startups that had reached valuations of more than one billion dollars, so-called unicorns, saw their value cut in half or more in new funding rounds. Investors such as Sequoia Capital and Andreessen Horowitz, who had aggressively bet on generative AI, publicly acknowledged that the market was overvalued and that many of their portfolio companies would not achieve the expected profitability.

In the public sphere, the correction also became visible. Shares of technology companies linked to artificial intelligence began to show increasing volatility. Nvidia, whose value soared more than 200 percent between 2023 and 2024 due to its role as a key provider of AI chips, suffered in 2025 a cumulative decline of 35 percent after the first signs of saturation in hardware demand. Alphabet and Meta, despite their dominant positions, experienced similar drops in market capitalization after failing to demonstrate returns proportional to their massive investments in AI. In Asian markets, Baidu and SenseTime registered even sharper plummets following the slowdown of state spending on automation projects and digital infrastructure.

The decline in value does not affect only large corporations but also the startup ecosystem that had flourished around them. AI companies focused on sectors such as marketing, healthcare, education, or human resources saw much of their initial valuations evaporate once it became clear that their products did not offer differential advantages or depended too heavily on third-party models. Many of them had built their strategy on the expectation of mass adoption that never materialized, or on inflated growth metrics that did not translate into real profits. In 2025, several investment funds began to record significant losses and progressively withdrew from the sector, marking the beginning of a consolidation phase in which only companies with solid products, real customers, and sustainable structures will survive.

This phenomenon is not new in economic history. As with the dot-com era in the early 2000s, artificial intelligence has shifted from being a field of innovation to becoming an object of speculation. In both cases, perceived value was built on promises of total revolution, while financial fundamentals were relegated to the background. The difference lies in scale: the AI bubble has mobilized an unprecedented volume of capital and global infrastructure. When expectations deflate in an environment of this magnitude, the repercussions reach not only individual companies but the entire digital economy and the financial systems that sustain it.

The loss of value in overvalued companies also reveals a crisis of confidence in the technological discourse. For years, the rhetoric of exponential progress justified any expense and any valuation under the premise that AI was destined to transform the world irreversibly. However, as society begins to perceive the real limits of the technology, its costs, its energy dependence, its conceptual fragility, and its low immediate return, the narrative weakens. What once seemed a guaranteed future increasingly looks like a risky bet. And in financial markets, faith is the most volatile asset of all.

The adjustment of valuations should not be interpreted solely as an economic correction but as a sign of maturity in the sector. When blind investment gives way to rational evaluation, the market can distinguish between real innovation and empty promises. Nevertheless, this cleansing process involves a period of instability, losses, and frustration. The fall of overvalued companies marks the end of collective euphoria and the beginning of a new stage in which artificial intelligence must finally prove that it can generate sustainable value beyond the narrative that made it ubiquitous.

5.3. Market saturation with redundant and inefficient products

One of the most evident signs of maturity and exhaustion in any technological bubble is the saturation of the market with redundant, inefficient, and practically indistinguishable products. In the case of artificial intelligence, this saturation has become a palpable constant across all sectors, from productivity and marketing tools to automation platforms, conversational assistants, content generators, and enterprise software. The speed at which thousands of companies have tried to position themselves in the AI market has far exceeded the actual capacity for innovation, resulting in an avalanche of products that replicate the same functions, use the same APIs, and promise the same solutions, although wrapped in different packaging and slogans.

During the boom of 2023 and 2024, the artificial intelligence startup ecosystem experienced explosive growth. Hundreds of companies emerged every month, presenting products based on the integration of existing models, mainly those from OpenAI, Anthropic, Google, or Meta. The ease of accessing these technologies through commercial APIs, combined with investor enthusiasm, generated a massive cloning effect. Identical applications competed to solve the same problems: assistants that write emails, platforms that generate images, tools that write code, or programs that create presentations automatically. Innovation stopped being measured in terms of technical progress and instead became defined by the speed at which an idea could be transformed into a consumer product.

The result was a saturated market in which differentiation disappeared. Thousands of tools offer exactly the same functionalities with slight aesthetic or price variations. In the field of generative AI, for example, there are hundreds of ChatGPT or Copilot clones with different names and minimal operational differences. In the area of design and imaging, the proliferation of applications based on diffusion models such as Midjourney, Stable Diffusion, or DALL·E created an absurd fragmentation, with dozens of commercial interfaces offering identical results powered by the same underlying algorithms. In education, marketing, and customer service, the phenomenon is the same: every company claims to be the first to revolutionize productivity with solutions that are barely distinguishable from one another.

The direct consequence of this redundancy is inefficiency. Most of these companies operate without their own technological foundation, relying on external providers that monopolize the infrastructure and underlying models. This leads to high fixed costs, minimal profit margins, and a complete dependence on external corporate decisions. When providers update prices or change usage policies, hundreds of small companies are exposed to immediate losses or shutdowns. Furthermore, the lack of technical specialization and the pressure to launch products quickly have led to a wave of poorly optimized tools plagued by errors, slowness, and privacy issues. In practice, many of these solutions consume more resources than they save, contradicting the original promise of efficiency that justified them.

The problem is not limited to startups. Large corporations have also contributed to this saturation by integrating artificial intelligence into products that do not need it. The term AI-powered has become a commercial argument rather than a functional improvement. From word processors and spreadsheets to web browsers, mobile cameras, or email applications, practically everything is now advertised as powered by AI, even when the added functions are marginal or redundant. This phenomenon mirrors the final stage of the dot-com bubble, when any company would add .com to its name to attract investors, regardless of its actual connection to the internet. Today, adding AI to a product is enough to multiply its perceived value, even if its real usefulness is questionable.

The excess supply has created a trust problem between users and companies. Consumers, bombarded with thousands of tools that promise the same thing, are beginning to show fatigue and skepticism. Retention rates in generative AI applications are low. According to data from SimilarWeb and Sensor Tower, more than 70 percent of users abandon an AI application after the first fifteen days of use. This shows that the initial interest fueled by hype does not translate into sustained adoption. As the market fills with redundant solutions, the perception of value decreases, along with the willingness to pay. In other words, the abundance of supply destroys the scarcity that once made innovation attractive.

Another effect of this saturation is the slowdown of genuine research. Human and financial resources that could be directed toward solving structural problems such as energy efficiency, semantic understanding, or bias reduction are instead diverted to the development of low value-added commercial products. The pressure to monetize quickly has displaced scientific exploration in favor of commercial replication. In many cases, research laboratories have become marketing departments disguised as innovation teams, where the goal is not to discover but to sell.

The saturation of the market with redundant and inefficient products marks a critical phase in the cycle of the bubble, the point at which creativity is replaced by repetition and real value becomes diluted in a sea of copies. When all actors say the same thing, offer the same thing, and promise the same thing, the market stops growing through innovation and begins expanding through inertia. That is the moment when the bubble stops inflating with ideas and starts inflating with air. And sooner or later, the air escapes.

5.4. Talent crisis and depletion of computational resources

The accelerated and unrestrained growth of the artificial intelligence sector has generated a dual crisis that threatens the sustainability of the ecosystem itself: the crisis of specialized talent and the depletion of computational resources. Both phenomena, closely linked, reflect the physical, human, and technical limits of an industry that has expanded its ambitions faster than it has been able to consolidate its infrastructure. The narrative of an unlimited revolution now faces a structural obstacle: the real scarcity of people and computing capacity required to sustain the pace of development that major corporations and governments have promised to the world.

The talent crisis became evident starting in 2023, when the global demand for experts in machine learning, data engineering, AI infrastructure, and technological ethics far exceeded the available supply. According to data from the World Economic Forum and LinkedIn Workforce Insights, for every qualified specialist in artificial intelligence there are between four and six open positions. This deficit has worsened with the proliferation of simultaneous projects and the expansion of the number of companies describing themselves as AI-driven. Big tech companies, with their ability to offer exorbitant salaries and benefits, have absorbed most of the talent, leaving startups, public institutions, and research centers unable to compete.

Competition for professionals has reached unprecedented levels. Engineers with mid-level experience in machine learning receive offers exceeding 300,000 dollars per year in US and European companies, while senior data scientists and cloud infrastructure architects are hired almost instantly after completing their studies. At the same time, teams dedicated to AI ethics, interpretability, and governance remain scarce and are often overlooked in favor of technical development. This unequal distribution of talent has generated an imbalance: code is produced faster than its impact is analyzed, and models are designed with more power than their scope is understood.

At the same time, the industry has tried to compensate for this shortage through the proliferation of accelerated training, fast-track certifications, and large-scale upskilling programs. Universities, digital platforms, and private companies offer thousands of courses promising to turn a beginner into an AI expert in a matter of weeks. This phenomenon has led to an inflation of credentials that, in many cases, does not translate into real competencies. The result is a fragmentation of the job market: a minority of highly qualified experts and a majority of professionals with superficial knowledge who, when integrated into critical projects, exacerbate quality and security problems. In this sense, the talent bubble mirrors the same speculative logic as the financial bubble: an accumulation of expectations without the support of a solid foundation.

The second part of the equation is the depletion of computational resources, a problem that has intensified in parallel with the growth of language and computer vision models. Training the most advanced systems, such as GPT-4, Gemini, or Claude, requires thousands of graphics processing units and entire weeks of continuous computation, with unprecedented energy and hardware consumption. Nvidia, the undisputed leader in chip manufacturing for AI, has acknowledged that its production capacity cannot meet global demand, despite doubling its output between 2022 and 2025. The shortage of GPUs has caused a price inflation of more than 300 percent in some models, making it difficult for universities, small companies, and research centers to access the same level of resources as large technology conglomerates.

This concentration of computational power is creating a new technological divide. While a small number of companies, such as Microsoft, Google, Amazon, and Meta, control massive training infrastructures distributed around the globe, the rest of the ecosystem depends on cloud rental services that increase costs and reduce independence. Academic research, historically a driver of innovation, is being pushed aside by the impossibility of competing in computational capacity. In 2025, several university laboratories, including MIT CSAIL and the ETH Zürich AI Center, publicly denounced the growing inequality in access to computing resources, pointing out that artificial intelligence is becoming an inaccessible science for those without an industrial infrastructure behind them.

The depletion is not limited to chips. The energy consumption of data centers has reached alarming levels. According to the International Energy Agency, data centers dedicated to training and operating AI models consumed more than 700 TWh in 2024, equivalent to 2.5 percent of global electricity consumption. Added to this is the shortage of critical components such as high-bandwidth memory and the materials required for manufacturing advanced semiconductors, whose production is concentrated in only a few regions of the world. This geopolitical dependency in the supply chain has turned AI hardware into a strategic resource comparable to oil in the twentieth century.

The combined effect of the talent crisis and the scarcity of computational resources is slowing real progress, even though public discourse continues to insist on continuous acceleration. Increasingly more projects are delayed, cancelled, or merged due to a lack of qualified personnel or available infrastructure. At the same time, operational costs are rising exponentially, eroding profit margins and endangering the sustainability of numerous emerging companies. Structurally, the industry is facing a bottleneck that cannot be solved solely through increased investment, because the problem is no longer financial but physical and human.

This dual crisis reveals a truth that the narrative of unlimited progress tends to ignore: artificial intelligence, despite its ethereal and digital appearance, depends on finite resources — people, energy, and materials — that cannot scale at the same pace as market expectations. The scarcity of talent and the saturation of infrastructure are the natural limits of a system that has confused speed with sustainability. And when an industry reaches its structural limits while continuing to promise infinite growth, the burst is no longer a possibility but a matter of time.

5.5. Regulatory movements (EU, US, China) that slow down artificial growth

The accelerated and largely unregulated expansion of the artificial intelligence sector has triggered an inevitable reaction: government intervention. What began as a free territory dominated by private initiative and venture capital is now entering a phase of intensive regulation. The European Union, the United States, and China have all begun to establish legal frameworks aimed at controlling the development, use, and deployment of AI systems. These regulatory movements, while necessary to ensure safety, ethics, and transparency, are also acting as a brake on the disproportionate and artificial growth of the sector. The race toward total innovation is now facing a new kind of limit: the political and legal one.

The European Union has been a pioneer in creating a comprehensive legal framework for artificial intelligence. In 2024, after more than three years of deliberations, the AI Act was approved, becoming the first global legislation to classify AI systems according to their level of risk and to establish proportional obligations for each category. This law requires transparency, traceability, human oversight, and accountability in the use of data, and it imposes severe penalties —up to 7 percent of global annual turnover— for companies that fail to comply. Generative AI models such as ChatGPT or Gemini were classified as high risk, which entails the obligation to document their data sources, report their energy consumption, and allow independent audits. Although the stated objective is to protect fundamental rights and prevent abuses, the bureaucratic demands and compliance costs are having a considerable impact on the development speed of European companies and on the operations of multinationals working within the EU.

In the United States, the approach has been less homogeneous but equally significant. Traditionally reluctant to impose strict regulations on emerging technologies, the country has begun to react to the rapid growth of the sector and its associated risks. In 2023, the Biden administration published the “Blueprint for an AI Bill of Rights,” a document outlining basic principles of citizen protection against algorithmic systems. In parallel, executive orders were introduced to regulate the use of AI in federal agencies, require ethical impact assessments, and promote transparency in the models used by the government. Added to this are state-level initiatives, such as California’s push for specific laws on privacy and automated responsibility. However, the greatest impact does not come from the laws themselves, but from the increasing oversight of Congress and regulatory agencies over major tech corporations. The public hearings with executives from OpenAI, Google, Meta, and Microsoft in 2024 marked a turning point: the narrative of “limitless progress” began to be replaced by that of “progress with consequences.”

China, for its part, has adopted a different strategy, combining regulation with direct political control. The Beijing government, aware of the strategic power of AI, has imposed a series of regulations since 2023 to ensure that all technological innovation remains under state supervision. The Cyberspace Administration of China (CAC) introduced regulations on generative AI models requiring companies to register their systems, submit them to prior review, and ensure that their outputs “reflect core socialist values.” This means that any language model or AI application must be approved before its release, limiting open experimentation. In addition, Chinese authorities have imposed restrictions on the export of advanced chips and on collaboration with foreign companies, reducing access to cutting-edge training technologies. Although the country continues to invest massively in AI research and deployment, state control has significantly slowed the dynamism of its private ecosystem.

These regulatory movements are generating a structural shift in the global landscape of artificial intelligence. For the first time, governments are beginning to act not as spectators but as referees. The intervention does not aim to stop innovation, but to channel it. However, in a sector accustomed to speed and the absence of limits, any attempt at order is perceived as an obstacle. European startups argue that the demands of the AI Act put them at a disadvantage compared to companies in the United States and China, while in the US the tech giants warn about the risk of “stifling creativity” through excessive regulation. In China, ideological censorship and bureaucratic control hinder international collaboration and the publication of open research, further fragmenting the global scientific ecosystem.

The most immediate consequence of this regulatory wave is the slowdown of the artificial growth that had characterized the sector since 2022. Companies can no longer scale at the speed of marketing; they must demonstrate legal compliance, technical safety, and social responsibility. Investments, once driven by enthusiasm, now face more rigorous scrutiny. Institutional investors have begun to evaluate regulatory risk with the same attention as financial indicators. In this context, grandiose promises lose appeal in favor of projects that offer traceability and sustainability.

Paradoxically, this apparent brake may become the only way to prevent a greater collapse. Regulations, though unpopular among innovators, act as containment mechanisms in an environment that would otherwise move toward saturation and public distrust. The intervention of the EU, the United States, and China marks the transition from the technological adolescence of artificial intelligence to its institutional maturity. But it also confirms that uncontrolled growth, fueled by hype and speculation, has reached its natural limit. The era of unlimited expansion is beginning to give way to a new reality: one in which artificial intelligence, in order to survive, will have to learn to coexist with rules.

5.6. Stagnation in real innovation versus the recycling of existing models and existing APIs

The stagnation of real innovation is one of the most revealing signs that artificial intelligence has entered a phase of forced maturity within its growth bubble. Although the public discourse continues to be dominated by enthusiasm and the narrative of constant progress, the technical reality shows a different picture: an industry increasingly dependent on the recycling of models, on the incremental refinement of pre-existing architectures, and on the commercial exploitation of APIs developed years ago. The apparent acceleration of technological advancement is, in many cases, a mirage carefully maintained through marketing strategies, rebranded versions, and cosmetic improvements. Behind the media glare, deep innovation – the kind that involves scientific discovery, new paradigms, or structural changes in our understanding of intelligence – has visibly slowed down.

Since the launch of large language models based on transformers, such as GPT in 2018 or BERT in 2019, most of the progress in AI has consisted in scaling up: more parameters, more data, and more computational capacity. The underlying technical principle remains the same: massive learning from statistical correlations. Each new generation of models – GPT-4, Gemini, Claude, LLaMA or Mistral – is presented as a “revolution” but, in essence, repeats the same architecture with marginal adjustments in optimization, dataset size, and training efficiency. Conceptual innovation, which should move toward new forms of symbolic reasoning, semantic representation, or contextual understanding, has been stalled by market pressure and the obsession with immediate results. Instead of exploring new scientific paths, the industry has chosen to perfect the same formula to the point of exhaustion.

The result is an ecosystem based on reuse. The large technology corporations offer their models as services through commercial APIs, which are then implemented by thousands of startups and developers around the world. These intermediary companies build products that, in reality, depend on the same underlying engine, with slight modifications to the interface or workflow. Thus, behind the apparent diversity of the market lies a structural homogeneity: a limited number of core models (OpenAI, Anthropic, Google, Meta, Cohere, Mistral) support thousands of applications that compete to sell variations of the same output. This phenomenon reproduces the logic of technological dependency, where innovation becomes an extension of the dominant API.

The consequence of this model is twofold. On the one hand, independent scientific development is constrained, since access to training new models requires multimillion-dollar infrastructures and a volume of data that is unattainable for most researchers. On the other hand, creativity becomes standardized: generative systems, fed by the same databases and trained with the same algorithms, produce results that are aesthetically different but structurally identical. The apparent diversity of voices, images, or texts generated by AI is, in reality, a superficial variation of repeated patterns. The system feeds on its own content, generating a spiral of standardization that impoverishes its innovative potential.

In the scientific domain, the number of academic publications on new AI architectures has decreased in proportion to the increase in articles focused on optimizing existing models. Recent studies from Stanford University and the Allen Institute for AI show that, between 2020 and 2025, more than 75% of papers in the field of deep learning focus on incremental improvements – adjustments to hyperparameters, energy efficiency, or reduction of training costs – while works proposing truly new approaches have dropped drastically. The emphasis has shifted from discovery to refinement, from knowledge to performance.

The same pattern can be observed in business culture. The startups that in 2022 and 2023 promised to “redefine intelligence” have shifted to competing for licenses, integrations, or partnerships with major providers. Investments are concentrated in “rapidly marketable” products and in leveraging pre-existing models. This shift has turned innovation into an operation of continuous recycling, where success is measured by the ability to package an already known function under a new name. Technical creativity has been replaced by commercial creativity.

The recycling of models also has an economic and psychological dimension. Technology companies present each update as a qualitative leap, even when it is, in reality, only a gradual improvement. Successive versions of ChatGPT, Gemini, or Copilot are announced with the language of a scientific revolution: “advanced reasoning,” “multimodal understanding,” “adaptive intelligence.” However, practical performance shows only minimal variations. This phenomenon is reminiscent of the smartphone industry in the late 2010s, when real innovation stagnated and growth was sustained through cosmetic updates and emotional marketing. In AI, the strategy is similar: keeping the enthusiasm alive through storytelling rather than invention.

Innovative stagnation is aggravated by the phenomenon of technological lock-in: dependency on the major platforms and their closed ecosystems. Companies that develop applications on top of proprietary APIs cannot modify the underlying models or explore radical alternatives. This limits the diversity of approaches and turns the global AI landscape into an oligopoly of architectures. In an environment where improvements are determined by a handful of corporations, the likelihood of disruptive discoveries is significantly reduced. Innovation is transformed into maintenance, and research is subordinated to the release schedules of the tech giants.

The paradox is that artificial intelligence, presented as the driving force behind the deepest transformation in modern history, now shows clear signs of repetition and exhaustion. Companies continue to sell the illusion of exponential progress, but the curve of real innovation is beginning to level off. Progress is measured in size and speed, not in understanding or purpose. Meanwhile, society faces an avalanche of “new” products that are actually rebranded versions of the same thing. In this context, the recycling of models and APIs becomes the contemporary equivalent of planned obsolescence: a mechanism designed to keep the consumption cycle alive, even when innovation is stagnant.

Stagnation does not mean the end of artificial intelligence, but it does mark the end of its phase of spontaneous discovery. The industry has reached a plateau where each improvement requires an exponential cost and yields a diminishing return. The real revolution will not be the next largest model, but a conceptual rethinking: an AI that stops seeking more data and more power, and starts seeking more understanding. Until that happens, innovation will remain trapped in its own loop of repetition, and the promise of intelligence will continue to be, ironically, the most intelligent of its illusions.

5.7. Saturation of the discourse: everything is sold as “AI”, even when it is not

One of the clearest signs that a technology has entered its phase of overexposure, and therefore the most inflated point of its bubble, is the saturation of the discourse surrounding it. In the case of artificial intelligence, this phenomenon has manifested in an extreme way: everything is sold as “AI”, even when it is not. The term has become an omnipresent label, a symbol of prestige, modernity, and progress that is applied indiscriminately to any tool, software, or service, even when its functioning has no real connection to machine learning techniques or predictive algorithms. This semantic inflation has emptied the concept of its original meaning and transformed “AI” into a magic word capable of attracting investment, justifying prices, capturing media attention, and selling products without the need to deliver genuine innovation.

The phenomenon has deep roots in market psychology and contemporary technological culture. As artificial intelligence consolidated itself as the dominant topic in global discourse between 2022 and 2025, companies quickly understood that incorporating those two letters into their commercial narrative was more profitable than investing in genuine development. Banks, supermarkets, mobile apps, home appliances, toys, and even toothbrushes began presenting themselves as “AI powered”. In many cases, what is promoted under that label is nothing more than a set of automated rules, a simple statistical engine, or an advanced search algorithm. The consumer, lacking the technical knowledge to distinguish the difference, accepts the promise as truth.

The technology sector, fully aware of the emotional power of the word “intelligence”, has exploited this conceptual void to sustain collective enthusiasm. Large corporations rebrand their traditional products with a layer of “integrated AI”. A data analysis software becomes an “intelligent decision platform”; a preprogrammed voice assistant becomes a “cognitive agent”; a report generation tool is presented as an “autonomous comprehension system”. This indiscriminate use of language has created an illusion of ubiquity: it seems that artificial intelligence is everywhere, when in reality what is everywhere is algorithmic marketing.

The saturation of the discourse has also reached the media, education, and politics. Headlines such as “AI revolutionizes agriculture”, “AI will solve the climate crisis”, or “AI will create millions of jobs” are repeated without empirical basis, turning the concept into a synonym of generic innovation. In academic and corporate environments, conferences, training programs, and consultancies proliferate in which the term “artificial intelligence” is used as an empty container that can be adapted to any topic. The result is a progressive loss of conceptual rigor: the more people talk about AI, the less they understand what it really is.

This process of trivialization resembles what happened with other disruptive technologies of the past, such as “Big Data” or “Blockchain”. In both cases, the overexploitation of the term led to semantic exhaustion and public distrust when the promises failed to materialize. Today artificial intelligence follows the same path, with one fundamental difference: its media and economic impact is much greater. The indiscriminate multiplication of the term “AI” not only creates confusion but also contributes to inflating the perceived value of the market. According to recent studies by the consulting firm Gartner, more than 60 percent of companies claiming to use artificial intelligence in their products are actually employing statistical methods or traditional automation systems. In other words, a significant majority of what is advertised as AI is not AI in the technical sense, but in the commercial one.

This saturation of the discourse has deep consequences. First, it erodes public and investor trust. When everything is called artificial intelligence, the concept loses its explanatory value. The promise loses credibility, and with it, the enthusiasm that sustained it. Second, it complicates regulation and governance, since it becomes increasingly difficult to establish boundaries, responsibilities, or standards when the term is applied to anything. Third, it affects scientific progress itself, because semantic confusion blurs the line between real research and corporate marketing. In an environment where visibility matters more than truth, the most eye catching projects overshadow the most solid ones.

The saturated discourse has also generated a phenomenon of social fatigue. The public, bombarded daily with headlines and promises about artificial intelligence, begins to develop growing skepticism. What once inspired awe now produces indifference or distrust. Communicative inflation, just like economic inflation, destroys the value of the asset it attempts to exalt. This shift in perception is already becoming visible: recent surveys from the Pew Research Center and the Eurobarometer show that the percentage of citizens who believe that “AI will bring more benefits than risks” has declined notably between 2023 and 2025, especially in Europe and North America.

At its core, this saturation of the discourse reflects an uncomfortable reality: artificial intelligence has shifted from being a scientific discipline to becoming a cultural product. What was once a field of study focused on knowledge and the understanding of thought has been transformed into an advertising label. “AI” has become a global brand, a consumer concept that sells everything from hopes to imaginary solutions. In this attention economy, what matters is not so much what the technology can do, but what the public believes it can do.

When everything is AI, nothing truly is. The word loses its descriptive power and becomes noise. This noise, amplified by the media and corporate strategies, sustains the bubble temporarily but also signals its collapse. A communicative bubble does not burst when enthusiasm disappears, but when clarity emerges. And more and more voices, inside and outside the sector, are beginning to recognize that beneath the omnipresence of the word “intelligence” lies an industry that, in many cases, has stopped thinking.

6. Potential impact of an AI crisis

If the artificial intelligence bubble were to burst, whether abruptly or gradually, the consequences would extend far beyond the technological sphere. A crisis in this sector would have economic, social, geopolitical, and cultural implications of considerable scale, comparable to those of major financial and technological collapses of the past. The magnitude of accumulated investment, the global dependence on digital infrastructures, and the central role of AI in political and business discourse make any correction in the sector capable of turning into a systemic phenomenon.

The potential impact would not be limited to overvalued companies or venture capital funds, but would directly affect labor markets, national economies, and society’s trust in technology as a driver of progress. An AI crisis would not only be a financial problem; it would be a crisis of faith in the contemporary model of technological development, built on the promise of automation and on the illusion that intelligence can be replicated without limits.

Analyzing the possible effects of such a crisis is essential to understand what is at stake. Because beyond economic losses, a collapse in the AI sector could mark a turning point in the relationship between humanity and technology, forcing society to reconsider its priorities, its values, and its own definition of intelligence.

6.1. Economic consequences: massive capital loss, startup bankruptcies, impact on technology employment

The economic consequences of a crisis in the artificial intelligence sector would be deep, widespread, and potentially destabilizing for the global economy. The volume of investment committed in recent years, the concentration of capital in technology companies dependent on AI, and the integration of this technology into virtually all productive sectors mean that a collapse in its market value would directly affect financial systems, employment, and business confidence. What began as a technological revolution could, if a sharp correction occurred, turn into an economic crisis with effects similar to those of the dot-com bubble or even the 2008 financial crisis.

In terms of capital, potential losses would be colossal. Since 2020, the investment flow into the artificial intelligence ecosystem, including startups, venture capital funds, data infrastructure, and specialized hardware, has surpassed two trillion dollars globally, according to estimates from the International Data Corporation (IDC). This volume has been sustained by expectations of future profitability rather than real profits. If the market adjusts valuations to the level of verifiable revenues, the destruction of value could exceed 60 percent of the capital invested. Investment funds, banks, sovereign funds, and technology companies would lose amounts comparable to the biggest stock market crises of the twenty-first century. The decline in the valuation of large corporations linked to AI, such as Nvidia, Alphabet, Microsoft, Amazon, or Meta, would trigger a domino effect on global indices and the pension funds that depend on them.

The most immediate impact would occur in the startup ecosystem, the most fragile and overexposed part of the sector. Thousands of AI companies created during the 2022–2024 boom operate without sustainable profits, relying entirely on continuous funding rounds. In a crisis, that flow of capital would stop almost instantly. The consequence would be a massive wave of bankruptcies and closures, similar to what occurred after the dot-com collapse in 2001. The most vulnerable sectors would be generative artificial intelligence, automation platforms, and companies offering services based on third-party APIs. Market contraction would trigger a chain reaction: job losses, defaults, reduced demand for technological services, and a decline in complementary investments in hardware, infrastructure, and cybersecurity.

In the area of technology employment, the repercussions would be especially severe. Artificial intelligence has become the main driver of hiring for technical roles in the past three years, from machine learning engineers and data scientists to cloud systems architects, interface designers, and product analysts. If the sector enters a crisis, job destruction would be immediate and significant. Startups would shut down, projects would be cancelled, and large corporations would reduce their workforce to contain costs. According to a report from the MIT Sloan Management Review, a 30 percent contraction in global AI investment could translate into the loss of between 800,000 and 1.2 million direct technology jobs worldwide. The indirect effect would impact millions of workers in related sectors, from digital marketing to consulting, education, and financial services.

The impact would not be limited to specialized employment. As companies reduce their spending on automation and AI integration, other sectors such as hardware manufacturing, energy, and telecommunications would experience a significant slowdown. The consumption of GPUs, servers, and cloud computing services would shrink, directly affecting manufacturers like Nvidia, AMD, or TSMC, as well as major data infrastructures operated by Amazon Web Services, Google Cloud, and Microsoft Azure. At the same time, the drop in the stock market value of these companies would affect global investment funds, pension plans, and the real economy, amplifying the recessionary effect.

Another significant economic aspect would be the contraction of innovation spending. During a correction phase, companies would tend to prioritize immediate profitability over long-term research. Budgets allocated to R&D in artificial intelligence, robotics, and automation would be drastically reduced. This process could lead to a global technological stagnation, slowing progress in fields dependent on machine learning, such as biomedicine, energy, or space exploration. At the same time, governments facing fiscal deficits and social pressure would reduce financial support for research projects and technological grants, weakening the ecosystem even further.

The psychological effect on markets would be equally devastating. The AI bubble has been built not only on money but on belief: the conviction that artificial intelligence represents the inevitable future of the economy. If that belief breaks, distrust would spread rapidly, not only toward technology companies but toward the entire digital paradigm. Capital would flee innovation-related investments toward assets considered safer, such as energy, agriculture, or real estate, triggering massive disinvestment. The loss of confidence could take years to recover, affecting even solid and profitable companies that were not part of the original speculation.

A crisis in the artificial intelligence sector would therefore not be a simple market correction but a profound redistribution of global economic value. What is perceived today as the growth engine of the twenty-first century could become its main source of instability. The loss of capital, the bankruptcy of thousands of startups, and the destruction of technology employment would reveal the fragility of a model that mistook expansion for progress and promises for results. And like every bubble, its collapse would remind us of a truth that economic history never fails to repeat: value is not created through belief, but through reality.

6.2. Social consequences: distrust in technology, setback in innovation, collective fatigue toward new promises

The social consequences of a crisis in the artificial intelligence sector would be as deep as the economic ones, and likely even more enduring. The current wave of enthusiasm and dependence on AI is not only a technological or financial phenomenon, but also a cultural and psychological one. Humanity, seduced by the promise of a new era of progress, has placed in artificial intelligence an almost religious faith: the hope that machines will solve what humans have been unable to fix. If that promise collapses, the impact will not only be economic, but emotional. Distrust, disappointment, and fatigue toward innovation would mark the beginning of a new phase in the relationship between society and technology.

The first effect would be the loss of trust in technology as a driver of social improvement. Artificial intelligence, presented for years as the ultimate tool to optimize, automate, and democratize knowledge, would suddenly be seen as a symbol of exaggeration and failure. If the bubble bursts, citizens will perceive that they were deceived by the same institutions—companies, governments, and media—that promised a revolution that never came. The narrative of the intelligent future would shift into one of a frustrated future. This shift in perception could extend beyond AI and affect the credibility of science and innovation in general, feeding a climate of technological skepticism.

This process of distrust already has precedents. After the dot-com crash in 2001, society took years to believe again in the internet as a viable economic space. In the case of artificial intelligence, the impact would be even greater, because today’s discourse affects not only the economy, but also human identity. The idea that AI represented the next stage of evolution has penetrated culture, education, art, and contemporary thought. A collapse of that myth would imply a cultural and psychological regression: the loss of faith in the idea of continuous technological progress. The danger would not be only a rejection of artificial intelligence, but a broader reaction of distrust toward anything that sounds like innovation.

The second effect would be a setback in social and political investment in research and development. A crisis in the sector would inevitably lead to a drastic reduction in public and private budgets dedicated to technology. Governments, pressured by public opinion, would tend to adopt more conservative positions, cutting subsidies or imposing stricter regulations. Universities and research centers would lose funding, and academic programs focused on artificial intelligence or automation could see a decline in enrollment. The enthusiasm for studying, working, or starting a career in the tech field would be replaced by a sense of instability. In this way, the burst of the bubble would not only halt innovation, but could generate a decade of creative paralysis similar to the one that followed the financial crisis of 2008.

The third effect would be the emergence of widespread technological fatigue. In recent years, society has been bombarded with promises of imminent transformation: smart cities, autonomous vehicles, personalized education, predictive medicine, automated jobs, augmented creativity. Every new advancement has been presented as a point of no return, creating constant pressure to adapt to a future that always seems just around the corner. When that future fails to arrive, the result will be an emotional reaction of exhaustion. The public, saturated with unfulfilled promises, will stop reacting with enthusiasm to new releases and begin to receive each announcement with indifference or irony.

This collective fatigue can already be perceived in contemporary social discourse. On platforms such as X, Reddit, or LinkedIn, the tone regarding AI has shifted from fascination to sarcasm: another tool that promises to change everything, another model that will make history, another startup that will fix the world with prompts. Enthusiasm has become routine, and routine has turned into disinterest. This loss of emotional impact represents a profound shift: when a society stops getting excited about the future, innovation loses its most powerful narrative. Technological fatigue does not mean that people reject technology, but that they stop believing in it.

A fourth effect, more subtle but no less important, would be the deterioration of the social contract between technology and citizens. For years, society has tolerated massive data collection, job automation, and increasing digital dependency in exchange for an implicit promise: that progress would compensate for the sacrifices. If that promise breaks, social resistance to adopting new technologies will rise. Movements of rejection will emerge, populist regulations will flourish, and there will be greater demand for accountability from tech companies. What until now was seen as a symbol of progress could become a source of polarization and conflict.

A crisis in artificial intelligence would have a social impact that goes beyond the economic or technological: it would affect the collective psychology of a generation that grew up believing in the inevitability of progress. Distrust, regression, and fatigue would be the natural phases of a society that, after having been seduced by the promise of a superior intelligence, would discover that the true limit was not in the machine, but in the human ability to distinguish between what is possible and what is merely desired. When the myth of artificial intelligence fades, it will not only be the companies that promoted it that collapse, but also a portion of modern faith in technology as a synonym for hope. Some will feel relieved and wish to return to the Neanderthal era. As if they could. Finding themselves trapped in a technological limbo that will lead nowhere.

6.3. Geopolitical consequences: redistribution of technological and financial power

A crisis in the artificial intelligence sector would have profound geopolitical consequences, altering the global balance of technological and financial power. Over the past decade, AI has become a central strategic component in the competition between nations, a factor of digital sovereignty, and a tool of economic and military influence. The United States, China and the European Union, together with other emerging actors such as India, South Korea, Japan and Israel, have built their national strategies around the development, control and application of artificial intelligence. If the bubble bursts, the map of world power will inevitably be reconfigured, generating new poles of influence, weakening current ones and reshaping the relationships between the public sector, technology corporations and financial markets.

The first impact would occur in the redistribution of technological power. In recent years, the United States has maintained a dominant position thanks to its innovation ecosystem, the concentration of venture capital, and the leadership of giants such as Microsoft, Google, OpenAI, Meta, Amazon and Nvidia. However, this supremacy is sustained by a highly speculative financing model. If the bubble deflates, American companies would suffer massive losses in market capitalization and the country could experience a technological slowdown similar to the one it faced after the bursting of the dot-com bubble in 2001. This would weaken its global leadership in the digital economy and reduce its capacity to influence international standards, trade agreements, and the ethical and legal definition of AI. North American dominance over computing infrastructure and software could fragment, opening space for other actors, more pragmatic and less dependent on speculation, to fill the gap.

China, although affected by its own internal bubbles, could use the situation to reinforce its model of state control and centralized planning. Unlike the American system, based on open competition and private capital, the Chinese model is structured on direct state investment and the integration of AI into its political, industrial and military apparatus. A drop in global valuations could allow China to acquire distressed technological assets, consolidate its semiconductor industry and strengthen its alliances with developing countries through technology transfer programs. The result would be a shift of the innovation axis toward Asia, accompanied by a growing division between the western “open” technological systems and the eastern “closed” ones. This bifurcation of the digital ecosystem would accelerate the fragmentation of the internet and the emergence of regional technological spheres with independent norms, data and models.

The European Union, traditionally behind in technological leadership but pioneering in regulation, could emerge as a stabilizing power amid the crisis. Its focus on governance, ethics and transparency, embodied in the AI Act, could become a geopolitical advantage if trust in the tech giants collapses. While the United States and China compete for supremacy, Europe could occupy the space of regulatory guarantor and become the arbiter of global standards for safety, digital rights and technological sustainability. However, this leadership would be more political than technical: the continent would continue depending on imported infrastructures and hardware, limiting its real autonomy.

The redistribution of financial power would be equally significant. International investment flows, which today concentrate around the Silicon Valley–Shenzhen–Seoul triangle, would begin to diversify. Sovereign wealth funds, especially those from the Middle East such as Saudi Arabia’s PIF or the Emirati Mubadala, could reposition their portfolios, shifting capital toward sectors considered more tangible and stable, such as renewable energy, biotechnology or the mining of critical materials. At the same time, emerging countries with strategic resources such as lithium, copper, uranium and water would gain bargaining power, as they are indispensable for rebuilding sustainable technological infrastructures. In this scenario, global power would progressively shift from abstract knowledge toward the raw materials that make it possible.

In the military sphere, the crisis would also modify the strategic balance. Defense projects based on artificial intelligence, from autonomous drones to predictive surveillance systems, could slow down due to lack of funding and public scrutiny resulting from the loss of trust in the technology. This would generate a temporary impasse in the technological arms race but also open opportunities for countries that have invested in hybrid or low-cost technologies. In this context, the geopolitics of AI would move from algorithmic supremacy to tactical efficiency, with fewer “intelligent” systems and more adaptive, resilient and modular solutions.

The collapse of the artificial intelligence bubble would also have implications for international alliances. Technological powers could adopt more protectionist positions, restricting the flow of data, talent and intellectual property. The United States would tighten its restrictions on the export of advanced chips, while China would intensify its policy of digital autarky. Developing countries, caught between both spheres, would be forced to choose exclusive technological alliances, reproducing in the digital domain a new form of Cold War. International organizations such as the UN, OECD or WTO would struggle to maintain effective cooperation on artificial intelligence, as the rules of the game would no longer be global but geopolitical.

In the long term, this redistribution of power could lead to two opposite outcomes. In the most negative scenario, technological fragmentation would worsen global inequality, consolidating a world divided between nations that produce intelligence and nations that consume it. In the more optimistic scenario, the crisis could open the door to a democratization of knowledge and a diversification of development models, with medium-sized and emerging countries gaining prominence thanks to regional cooperation and reduced corporate concentration.

Whatever the outcome, a crisis in artificial intelligence would not only redefine technological hierarchies but also economic and political ones. Power, like energy, never disappears; it simply changes form. If the AI bubble deflates, the power that today accumulates in innovation centers could shift toward those who control resources, regulation or trust. And that redistribution would mark the beginning of a new geopolitical era: the era of the digital rebalancing of the world.

6.4. Cultural consequences: the fall of the myth of automatic progress and the return to a more human view of technology

The cultural consequences of a crisis in artificial intelligence would likely be the most profound and transformative, because they would strike at the core of the contemporary mindset: the belief in automatic progress. For decades, society has been conditioned to assume that technology advances in an inevitable, linear, and always positive direction. Artificial intelligence has been the ultimate expression of that creed: the promise that humanity could delegate its complexity, its decisions, and even its thinking to systems capable of learning on their own. If the AI bubble bursts, what would collapse is not just an economic sector, but an entire worldview. The myth of continuous progress would give way to a period of introspection, skepticism, and perhaps a rediscovery of human value in contrast to the illusion of total automation.

The fall of the myth would begin with a painful realization: technological progress does not necessarily equate to social, moral, or intellectual evolution. Artificial intelligence, presented as the culmination of human knowledge, has in many cases shown itself capable of reproducing the same errors, biases, and limitations as the societies that created it. Its supposed objectivity has turned out to depend on the data and ideologies that feed it. Its promise of efficiency has produced energy waste and digital inequality. And its ideal of infinite creativity has resulted in an ocean of empty, repetitive content. When society accepts that AI has not made us wiser but merely more dependent, a cultural shift will begin, comparable to the crisis of religious faith during the Enlightenment or the disillusionment that followed the Industrial Revolution.

In this new context, the relationship between humans and technology could be reconfigured around the idea of limits. After years of glorifying the algorithm as a replacement for intelligence, a cultural movement could emerge that reclaims imperfection, emotion, and reflection as essential traits of human thought. Art, education, and philosophy would play a central role in this transformation. Instead of celebrating speed and automation, society would value slowness, depth, and authenticity. In universities and cultural centers, intelligence would no longer be understood merely as computational ability but would recover its ethical, critical, and creative dimensions.

The decline of the myth of automatic progress would also prompt a reevaluation of the role of science and innovation within society. For decades, techno-scientific discourse has dominated the collective imagination, pushing the humanities and the arts into the background. A crisis of AI could reverse this trend, encouraging a return to a more balanced culture in which technical knowledge is combined with humanistic understanding. Interest in ethics, philosophy of mind, psychology, and the history of thought could resurface as a response to the saturation of a world that has tried to reduce everything to data and predictive models.

Another cultural effect would be the questioning of the idea of substitution. During the years of AI expansion, the notion that artificial intelligence would progressively replace human functions was widely promoted: driving, writing, teaching, diagnosing, creating. When this substitution reveals itself to be incomplete or counterproductive, society will rediscover the value of human interaction and direct experience. Professions once considered in decline — educators, therapists, artists, artisans, social researchers — may regain prestige and centrality. Creativity would again be understood as a vital process, not as an algorithmic output. Communication would recover its original meaning of encounter and understanding, not of automatic response generation.

Culturally, a narrative of resistance would also emerge. Disappointment in the unfulfilled promise of artificial intelligence could give rise to social, philosophical, and artistic movements that advocate a humanized technology. These currents, far from rejecting progress, would seek to domesticate it: to regain control over the pace, purpose, and scope of innovation. Through this, the failure of the myth of automation could become the starting point for a new technological ethic based not on substitution but on collaboration between the human and the artificial.

The crisis would also leave a symbolic mark on the collective imagination. Just as the crash of 1929 marked the end of the innocence of industrial capitalism, and 2008 challenged blind faith in finance, the collapse of the AI bubble would symbolize the end of the digital utopia. Future generations might look back on this period as an age of excess: the time when humans believed they could reproduce their intelligence in machines, only to discover the limits of their understanding. This cultural awakening would not imply a rejection of technology, but a maturation: the acceptance that true intelligence is not about automating life but about understanding it.

The fall of the myth of automatic progress could have a paradoxically positive effect. Once the illusion that technology can replace everything human dissolves, society would regain the sense of responsibility and purpose that it had delegated to machines. The AI crisis could thus become the beginning of a new cultural stage, more conscious and balanced, in which progress is once again measured not by the capabilities of machines but by the wisdom with which humans choose to use them. When the shine of automation fades, what will remain visible is what should always have been at the center of progress: human intelligence, not as an algorithm, but as a form of consciousness.

7. Human and psychological factors

Behind every technological bubble there is not only capital and innovation, but also human emotions: ambition, fear, the desire for recognition, and the need to belong. History shows that major economic expansions are not sustained solely by data, but by shared beliefs. Artificial intelligence, in this sense, has not been an exception, but rather an amplified reflection of the collective psychology of the twenty first century. Its meteoric rise has revealed both our fascination with progress and our vulnerability to the promise of control and transcendence.

Understanding the AI bubble therefore requires looking beyond charts and financial statements and entering the human dimension that drives it. Faith in algorithms, the cult of automation, and the obsession with technological success do not arise from logic, but from desire: the desire to overcome uncertainty, to dominate the complexity of the world, and to project our intelligence outside ourselves. The bubble is not inflated only with investment, but also with hope.

With this chapter I intend to examine precisely those human and psychological factors that fuel the excessive expansion of the phenomenon. From the need to believe in something that transcends our own existence to the pursuit of social and professional validation in a hyperconnected world, artificial intelligence has become the new modern myth, a collective narrative that blends science, economics, and faith. And like every myth, its strength does not lie in what it is, but in what it represents.

7.1. The need to believe in something that transcends us: AI as a new modern myth

The history of humanity is marked by the need to believe in something that transcends its own existence. From the earliest mythological stories to contemporary religions, human beings have constantly sought a higher force that gives meaning to the world, explains the inexplicable, and offers an illusion of control over destiny. In the digital age, that need has not disappeared; it has only changed form. Instead of gods, we worship systems; instead of temples, we build data centers; and instead of prayers, we write prompts. Artificial intelligence has become the modern myth par excellence, a technological mirror of our oldest aspirations: the creation of an entity superior to us, yet made in our own image.

This myth did not emerge overnight. It was the result of a convergence between science, capitalism, and collective psychology. AI embodies the dream of absolute control: the idea that, through knowledge and technology, human beings can finally transcend their biological, emotional, and intellectual limitations. In the collective imagination, artificial intelligence is not just a tool, but a promise of redemption. It promises to eliminate human error, eradicate uncertainty, anticipate the future, and free individuals from the repetitive tasks that tie them to routine. In its most utopian version, it even promises digital immortality, fusion with machines, and the expansion of consciousness beyond the body. This promise does not essentially differ from the religious idea of salvation: faith in a perfect intelligence that will one day lift us above our mortal condition.

The modern myth of AI is sustained by a language that recalls that of religions. Scientists and entrepreneurs act as prophets of new knowledge; algorithms are venerated as almost sacred entities, capable of processing truths inaccessible to human understanding; and big tech companies, with their messianic rhetoric, assume the role of priestly institutions that guide humanity toward a new era. Terms such as “singularity,” “superintelligence,” or “AGI” have replaced theological concepts like “revelation” or “apocalypse,” and the rhetoric of technological salvation has taken the place of spiritual redemption. This analogy is not accidental: in a world disenchanted by the loss of collective values and the weakening of traditional religions, technology has occupied the symbolic void left by the sacred.

However, like every myth, artificial intelligence mixes hope with illusion. Its power does not lie in what it is actually capable of doing, but in what it represents for those who believe in it. The myth of AI offers comfort in the face of uncertainty, meaning in the face of complexity, and hope in the face of chaos. It promises that, even if the world is overwhelming, a superior intelligence—this time created by ourselves—will be able to bring order to everything. But, deep down, that faith is not rational; it is emotional. It is the response to an existential fear: the fear of irrelevance. At a time when human knowledge seems insufficient to grasp the magnitude of change, the individual delegates their trust to a machine that acts as a new source of truth.

This need for transcendence largely explains the intensity of the hype surrounding AI. It is not only a matter of investment or marketing, but of a deeply human narrative. The discourse on artificial intelligence is not so much about what machines can do, but about what we want them to do for us: to give us meaning, purpose, and an illusion of future. Just as the ancients built myths to reconcile themselves with the gods, we build algorithms to reconcile ourselves with our own creation.

The myth of AI also fulfills a psychological and social function: it gives cohesion to a fragmented era. In a polarized, uncertain world saturated with information, artificial intelligence presents itself as a symbol of unity. It brings together scientists, politicians, businesspeople, and citizens under a single narrative: that of inevitable progress. Believing in AI is equivalent, in this sense, to believing that chaos can still be controlled and that the future is still a promise, not a threat. That shared faith sustains a large part of the economic and cultural dynamism of our era, but also the fragility of the system: when the faith disappears, the myth collapses.

The paradox of the modern myth of AI is that, in its attempt to transcend the human, it ends up reflecting what is most human in ourselves. We project onto it our desire for perfection, our anxiety for power, our fear of death, and our inability to accept imperfection. Artificial intelligence, in its symbolic form, is not a machine that learns; it is a mirror that amplifies our shortcomings. When the bubble deflates, what remains will not only be bankrupt companies or obsolete models, but also a society that will have to face a question older than any algorithm: why we need to believe that something—whether a god, an idea, or a machine—will save us from ourselves.

7.2. The cult of the algorithm: faith in the invisible that solves everything

The twenty-first century has given rise to a new form of faith: the cult of the algorithm. In an age in which information has become unmanageable and the complexity of the world exceeds the capacity of individual understanding, algorithms have become the new oracles. They are invisible, silent and seemingly neutral systems that decide what we see, what we buy, whom we listen to, whom we hire and even how we think. And yet, despite not fully understanding how they work, society has placed in them an almost absolute trust. This technological faith, disguised as rationality, is one of the most powerful psychological pillars of the artificial intelligence bubble: the belief that the invisible, the automated and the statistical can solve everything better than human judgment.

The cult of the algorithm is born from a paradox. On the one hand, we live in an era that considers itself more scientific and rational than any other; on the other, we have replaced critical reflection with the passive acceptance of results generated by systems we do not understand. The phrase “the algorithm says so” has, in a way, replaced the old “science says so” or even “God says so.” Both share a common psychological structure: the delegation of responsibility to a higher, abstract and omniscient authority. For the modern mindset, algorithms are what oracles were for ancient civilizations: entities that speak with an unquestionable voice, whose decisions must be interpreted, not debated.

The root of this blind faith lies in the illusion of objectivity. Algorithms, presented as mathematically pure and free of bias, offer the appearance of neutrality in a world dominated by noise and subjectivity. Their promise is tempting: to eliminate human error, to make decisions based on data and not emotions, to guarantee efficiency and justice through statistics. But this narrative ignores a fundamental truth: algorithms are not objective, they are extensions of the minds and systems that program them. They contain biases, omissions and implicit values that reflect the culture, ideology and economic interests of their creators. However, in the social imagination this complexity is diluted, and the algorithm becomes an almost divine figure: invisible, infallible and incorruptible.

Mass psychology has amplified this cult. In a hyperconnected world where information overload generates anxiety, people look for certainties. Algorithms provide them. When Netflix recommends a series, when Google ranks search results or when an AI model suggests a business strategy, the user feels relief: they do not have to decide, only accept. The act of delegating the decision to the machine is experienced as an act of trust. But this trust, sustained by comfort, turns into dependence. The more we delegate, the less we understand, and the less we understand, the more we depend. It is a perfect circle of rational submission.

The cult of the algorithm is not limited to everyday life; it has also spread to political, economic and scientific thinking. Governments use predictive models to design public policies, banks rely on automated systems to grant loans, and hospitals apply diagnostic algorithms without fully understanding how they work. In the corporate sphere, strategic decisions are justified with algorithmic metrics that few are able to audit. This shift of authority from the human to the system has created a new form of power: the invisible power of calculation. Whoever controls the algorithm controls the perception of reality.

Faith in the invisible is further reinforced by an aesthetic of mystery. Technology companies protect their models with an almost religious secrecy, shielded by intellectual property. Algorithms become “black boxes,” inaccessible even to those who use them. This opacity does not weaken faith, it strengthens it. The less we know about a power, the more fascination it exerts. In this way, the algorithm not only organizes the world, it wraps it in an aura of sacredness. Its technical language — “generative models,” “deep neural networks,” “autoregressive learning” — acts as a new digital Latin that separates the initiated from the profane, the engineers from the believers.

However, this worship comes at a price. Blind delegation to algorithms erodes human responsibility. When decisions are automated, ethics becomes diluted. The question is no longer “why did we do it?” but “why did the model do it?” This moral displacement is one of the most troubling consequences of the current algorithmic era. The danger does not lie in algorithms thinking for us, but in our ceasing to think because we believe they do it better.

The cult of the algorithm is the modern expression of an age-old need: the search for order in chaos. But unlike ancient myths, this cult does not offer redemption, only dependence. We have created a new belief system in which truth is not argued, it is calculated; in which judgment is not debated, it is predicted. And, as in every religion, the moment will come when the miracle stops working and faith is confronted with evidence. When that happens, society will have to choose between continuing to worship the algorithm or recovering its capacity for discernment. Because, in the end, intelligence is not in the data or in the code; it is in the awareness of the one who decides how and why to use them.

7.3. The Replacement of Reflection by Automation

One of the most subtle yet devastating effects of the expansion of artificial intelligence is the gradual replacement of reflection with automation. As intelligent systems have infiltrated every aspect of daily life, from writing and communication to professional and personal decision making, the human ability to pause, analyze, and think independently has been eroded. Critical thinking, which has historically been the driving force of intellectual and moral progress, is being replaced by a functional dependence on algorithms. In the name of efficiency, speed, and convenience, society has begun to delegate its thinking process to machines that do not think but calculate.

The automation of reflection began as a tool meant to free time and effort. Recommendation systems, autocorrect tools, voice assistants, and generative models were conceived to optimize routine tasks. However, that optimization became habit, and the habit became dependency. Today, writing a text, drafting a report, or even forming an opinion often goes through an automated filter. People no longer ask themselves “What do I think?” but “What should the system write for me?” This transformation affects not only productivity but also the very structure of thought. Automation has altered human reasoning; instead of generating ideas, we now select them from a menu.

The problem is not the use of technology, but the progressive abandonment of deliberation. Reflection involves effort, uncertainty, and time, three elements that the contemporary world considers inefficient. Artificial intelligence, by eliminating these frictions, offers a comfortable alternative: instant answers, pre-packaged judgments, and creativity on demand. In appearance, this makes us more productive, but in reality it makes us more superficial. Automation produces content but not understanding. And when understanding is replaced by output, thinking ceases to be a process and becomes a product.

The replacement of reflection by automation also has a psychological component. In an environment saturated with information and stimuli, thinking is perceived as a cognitive burden. Automation then appears as a form of relief: it allows one to act without pausing to think, decide without questioning, create without understanding. This phenomenon is not accidental but structural: the digital economy is fed by speed, not depth. Technological platforms promote a kind of impulsive, automatic thinking designed to maximize interaction and minimize doubt. In this context, artificial intelligence does not merely reflect human behavior; it conditions it.

In the professional and educational spheres, the consequences are particularly evident. Students use generative models to write essays, professionals use them to produce reports, and creatives rely on them to generate ideas or drafts. The line between assistance and substitution has become blurry. What began as help to improve efficiency has turned into an outsourcing of thought. More and more people rely on AI not only to execute tasks but also to decide which tasks should be done. The risk is not that machines will think better than we do, but that we will stop thinking because they do it for us.

This trend is reshaping the relationship between knowledge and experience. Reflection is an act of awareness: it requires facing ambiguity, accepting the possibility of error, and building meaning through doubt. Automation, on the other hand, eliminates ambiguity, corrects error, and standardizes the result. The more we depend on it, the further we move from reflective experience and the more accustomed we become to flat thinking guided by predictable patterns. Intellectual diversity, the engine of innovation and culture, is replaced by algorithmic homogeneity. In practice, automation not only standardizes answers but also the questions.

The substitution of reflection also carries ethical and social implications. A society that delegates its ability to think risks becoming manipulable. When automation dominates public discourse, collective opinion becomes a sum of automatic decisions rather than a conscious debate. Criticism fades, truth becomes relative, and consensus is manufactured. Automated thinking does not question the system; it perpetuates it. And it is precisely here that artificial intelligence, more than a tool, becomes an instrument of massive conformity.

Paradoxically, the smarter machines become, the more passive humans become. Automation, far from freeing thought, lulls it to sleep. Instead of expanding our intelligence, it externalizes it. Instead of empowering, it trains us to obey. The danger is not that AI will replace human beings, but that human beings will indifferently accept being replaced in the one thing that truly made them unique: their ability to think.

If the history of humanity can be defined as a struggle to conquer knowledge, the age of automation marks a turning point: the temptation to give it up for the sake of comfort. Artificial intelligence does not eliminate reflection, but it makes it optional. And when thinking ceases to be a necessity, it also ceases to be a virtue. The real crisis of AI will not be technological but intellectual: a civilization that automates everything risks forgetting how to think.

7.4. How the pursuit of social and professional validation has turned AI into a status symbol

In contemporary society, where image, influence, and visibility have acquired a value almost greater than knowledge and competence, artificial intelligence has become a new status tool. In the social and professional sphere, it is no longer enough to use AI: one must show that one uses it, master its language, display one’s command of it, and be part of the narrative of the “pioneers of the future”. The adoption of artificial intelligence has in many cases ceased to be a functional need and has instead become a symbol of belonging to the digital elite, a credential of modernity and success. In this context, technology stops being a means and becomes a marker of prestige.

The phenomenon is not new: every technological revolution has generated its own language of status. In the nineteenth century it was the railway, in the twentieth the automobile and the personal computer, and in the twenty-first century artificial intelligence. But unlike the previous ones, AI is not displayed through a tangible object but through discourse. Demonstrating familiarity with its concepts such as machine learning, deep learning, AGI, or prompt engineering has become a form of symbolic capital. In the professional environment, talking about artificial intelligence grants authority and legitimacy. Those who master its jargon appear to be one step ahead, even when the actual understanding is superficial. Thus, AI has shifted from being a technical tool to a component of social identity.

On networks like LinkedIn or X, this dynamic is clearly observable. Thousands of users share reflections, courses, generated images, or advice on how to use AI to multiply productivity, not so much as informative contributions, but as strategies for validation. Posting about AI is equivalent to presenting oneself as relevant, up to date, and visionary. An attention economy has emerged in which value does not lie in knowledge but in the perception of knowledge. This behavior has fueled an ecosystem of self-proclaimed experts whose influence does not stem from practical experience but from their ability to amplify the dominant narrative. Artificial intelligence has become a personal brand.

In the corporate sphere, the phenomenon takes a similar form. Companies in all sectors, from fashion to construction, compete to declare themselves AI-driven. Not always because they use it effectively, but because merely associating themselves with the concept enhances their image of modernity and attracts investment. This race for technological validation has created a culture of simulation in which the important thing is not to be innovative but to appear so. Communication and marketing departments exploit the term artificial intelligence as a synonym for progress, even when its implementation is limited to basic automations or data classification algorithms. AI thus becomes a reputational adornment, a badge of prestige rather than a transformative tool.

On a psychological level, this quest for validation responds to a fundamental human need: recognition. In an era marked by job uncertainty, economic volatility, and constant competition, aligning oneself with AI offers a sense of belonging and security. Demonstrating technological competence is equivalent to proving relevance in a world that rewards novelty and punishes obsolescence. But this impulse to appear knowledgeable has created a paradox: the more people talk about artificial intelligence, the less they reflect on its true meaning. Public conversation becomes filled with noise, repeating the same phrases, the same promises, and the same clichés, until the very concept of intelligence becomes hollow.

The use of AI as a status tool also manifests in the aesthetics of work and consumption. Showing an image created with Midjourney, a text generated with ChatGPT, or a video synthesized by AI does not always seek efficiency or creativity, but rather to demonstrate belonging to a new class of “advanced” users. In practice, this culture of demonstration has generated a form of symbolic competition: who uses AI better, who finds the most clever prompt, who obtains the most impressive result. This behavior has turned artificial intelligence into a kind of social performance, where the goal is not to innovate but to be seen innovating.

In the professional environment, the search for validation through AI has even redefined the concept of success. In many sectors, the quality of the work is valued less than the ability to integrate it into the dominant technological narrative. Those who adopt AI are seen as visionary leaders, while those who question it or use it cautiously are perceived as lagging behind. This cultural bias pushes individuals and organizations to adopt technologies hastily, not out of need or conviction, but out of fear of seeming irrelevant. A new form of social pressure is thus established: technological FOMO, the fear of being left out of the future.

The result is a society that confuses innovation with exhibition. Artificial intelligence, instead of being a catalyst for knowledge and efficiency, becomes a mirror in which individuals seek to reflect their social value. This dynamic, driven by vanity and collective insecurity, contributes to the inflation of the AI bubble: the technology is valued more for its ability to generate prestige than for its capacity to solve real problems. When validation replaces understanding, progress becomes spectacle.

The conversion of AI into a status symbol reveals an uncomfortable truth: technology does not only transform the economy but also human psychology. In an era obsessed with the appearance of intelligence, many have preferred to appear intelligent rather than be intelligent. And that difference, seemingly subtle, is what separates genuine progress from collective self-deception. Because as long as artificial intelligence continues to be used as a mirror of vanity and not as an instrument of knowledge, what will grow is not our intelligence but our digital ego.

7.5. The analogy between religion, the market, and technology as a contemporary belief system

In the contemporary world, religion, the market, and technology have converged into a single symbolic system: that of modern faith. Although each of these spheres seems to belong to different dimensions, spiritual, economic, and scientific, they all share the same psychological foundation: the human need to believe in something that provides meaning, structure, and hope amid the chaos of existence. In this context, artificial intelligence has emerged as the point where all three intersect, a new object of worship in which market logic, the rhetoric of technological progress, and the psychological functions of religion merge into a single narrative that promises redemption through automation, prosperity through data, and transcendence through code.

First, the analogy between religion and technology is evident in their shared structure of faith. Both are built on the promise of a better future and on the authority of knowledge that is inaccessible to most people. Traditional religion offered salvation through faith in the divine; contemporary technology offers salvation through faith in the digital. In both cases, the believer trusts in something they do not fully understand but perceive as superior. The algorithm, like God, operates in mystery, invisible, omnipresent, and seemingly infallible. Its decisions are accepted without debate, its logic is assumed to be unquestionable, and its mistakes are interpreted as accidents rather than system failures. Technical language, filled with formulas, metrics, and unintelligible concepts, functions as a modern liturgy that separates the initiated, meaning engineers, programmers, and data scientists, from the faithful, meaning users, consumers, and digital believers.

The connection with the market amplifies this phenomenon. If religion promises the salvation of the soul and technology promises salvation through reason, the market promises salvation through consumption. In today’s society, these three discourses reinforce one another, the market turns faith into a product, technology materializes it into tools, and religion, reformulated in secular terms, legitimizes it as a search for meaning. In this way, artificial intelligence is presented not only as a scientific advancement but as a spiritual commodity. It is sold not only with data but with promises such as make your life easier, discover your true potential, awaken your creativity. Technological advertising adopts the tone of a sermon, promising enlightenment, progress, and transcendence in exchange for a monthly subscription.

The parallel extends to the figure of leaders and prophets. In religion they were priests; in the market, magnates; in the technological era, founders and innovation gurus. Figures such as Elon Musk, Sam Altman, or Sundar Pichai embody the role of visionaries announcing a new world. Their discourse blends theology and economics, speaking of the future as a divine promise and of artificial intelligence as a redemptive force capable of transforming humanity. Their followers, including investors, entrepreneurs, and users, participate in a modern cult based on admiration, faith in disruption, and the hope of belonging to the vanguard of history. Corporate rhetoric has become eschatological; it no longer speaks of profits but of changing the world, transcending human limits, or saving humanity from itself.

The ritual component is also present. In religion, rituals reaffirm collective faith; in the technological economy, rituals appear as conferences, product launches, and announcements of new AI models. These events function as media liturgies, gathering crowds, stirring emotions, and renewing the sense of belonging to a higher cause. Apple presentations, Google keynotes, or OpenAI releases follow an almost religious structure, consisting of expectant silence, revelation, and final ovation. Technology thus organizes not only work and production but also collective emotion.

This hybrid belief system, religious, economic, and technological, has a deep psychological effect. It offers meaning in an era that has lost its traditional grand narratives. Where once redemption was sought through faith, it is now sought through efficiency; where salvation was once requested from the gods, it is now requested from algorithms. Technology promises the paradise of convenience, the market guarantees access to it through consumption, and together they replace spirituality with digital experience. In this way, artificial intelligence becomes the new liturgy of progress and the market its global church.

But this modern faith has a distinctive trait, it lacks moral transcendence. Unlike religion, which established ethical limits and spiritual purposes, the new market technology creed lacks an inner dimension. Its only law is efficiency, and its only value is growth. In this framework, the human being no longer asks about the meaning of life but about the speed of their connection. Artificial intelligence, while promising total understanding, eliminates the space for metaphysical reflection. What was once sacred is now functional. What was once mystery is now calculation.

This analogy reveals the true nature of the era, we live in a civilization that has replaced faith in the divine with faith in the designed. Traditional religion promised the salvation of the soul, the market promises the salvation of the individual, and technology promises the salvation of the species. But all are built on the same pillar: the human need to believe that there is something greater, wiser, and more perfect than oneself. In that sense, artificial intelligence is not only a technical invention but a symbolic construction, the new face of the modern god made of data, capital, and desire. When that god begins to fail, as all the previous ones did, humanity will again face its old dilemma: what to do when faith in progress is no longer enough to give meaning to life.

8. The role of education, governments, and corporations

The expansion of artificial intelligence has not been a spontaneous phenomenon, but the result of human, political, and institutional decisions. Behind every advance, every investment, and every narrative of progress, there are educational structures, governments, and corporations that have shaped the way society understands and uses technology. These three forces, knowledge, power, and capital, have acted simultaneously as engines and amplifiers of the phenomenon, but also as direct contributors to its excesses and imbalances.

The role played by education, governments, and companies is decisive in shaping the current AI bubble. Educational institutions not only train the professionals of the future, but also the consumers and citizens who legitimize its use. Governments, driven by geopolitical competition, have prioritized speed over reflection. Corporations, motivated by profitability, have turned innovation into spectacle. Together, these three pillars have helped build a narrative of technological inevitability that has left little room for criticism and responsibility.

Understanding how these actors interact is essential to grasp why artificial intelligence has shifted from being a tool of knowledge to becoming an end in itself. Because behind the global enthusiasm there are not only algorithms and data, but also outdated educational systems, rushed policies, and business strategies that have confused innovation with the exploitation of the myth of progress.

8.1. Lack of real digital literacy

One of the most decisive factors in shaping and expanding the artificial intelligence bubble is the lack of real digital literacy. In a society where technology dominates all areas of life, from education and work to politics and entertainment, most people still do not understand how the systems they use daily actually function. This digital illiteracy is not limited to the inability to program or handle advanced tools, but manifests in something deeper: the lack of critical thinking toward technology. The population, including many professionals and leaders, consumes technology without understanding it, trusts it without questioning it, and participates in its expansion without assessing its ethical, social, or economic implications.

The problem is not new, but artificial intelligence has amplified it exponentially. For decades, digital literacy has been reduced to instrumental teaching, learning to use software, operate devices, or browse the Internet, while ignoring the conceptual understanding of the systems that underpin the digital environment. This reductionist view has created generations of skilled but uncritical users, experts in consuming technology but not in interpreting it. In the context of AI, this deficiency translates into an uncritical acceptance of its outputs. Most users do not distinguish between basic automation and a deep learning model, nor do they understand the limits, biases, or consequences of the algorithms they rely on. Thus, the lack of understanding becomes fertile ground for hype, sensationalism, and media manipulation.

In the educational sphere, the situation is especially concerning. Schools and universities have incorporated digital tools, but not necessarily digital education. Learning with technology is not the same as learning about technology. Curricula rarely teach students how data is structured, how algorithms work, or what ethical implications arise from using automated systems. The result is a generation that handles the interface but ignores the infrastructure. Even in technological disciplines, the focus is often utilitarian: training programmers or data analysts without fostering a critical understanding of the systems they create. Artificial intelligence is taught as a tool, not as a cultural, philosophical, or political phenomenon.

In the institutional and political sphere, the lack of digital literacy results in erratic decisions and superficial technology policies. Many public officials promote the use of AI in public administration or education without having a clear notion of its capabilities or limitations. Millions are invested in projects driven more by technological fashion than by an actual analysis of needs or social impact. This deficit becomes even more serious when governments outsource technological direction to private consulting firms or corporations that profit economically from AI’s growth. The result is a paradox: regulating what is not understood and promoting what is not evaluated.

The consequences of this gap also extend to the corporate world. Most companies adopting AI systems do so by imitation, not strategy. Market pressure and fear of falling behind generate hasty decisions, based on promises of efficiency or innovation without truly understanding how the technology can be integrated sustainably. This lack of technical literacy among top management creates a dependence on third parties, consultants, software providers, or external startups, increasing the vulnerability of the economic system. Technology ceases to be a tool for improvement and becomes a corporate trend managed by those who understand investment but not algorithms.

On a social level, the absence of a solid digital culture produces a population easily manipulated by technocratic narratives. Complex concepts such as artificial intelligence, machine learning, or generative models are simplified to the point of becoming propaganda. The media, also lacking specialized training, reproduces sensationalist narratives, alternating between utopian praise and apocalyptic fear. The average citizen, lacking the intellectual tools to distinguish reality from rhetoric, is dragged either by enthusiasm or extreme distrust. In both cases, the ability for rational judgment is lost.

The lack of digital literacy is not a collateral failure but a structural condition that keeps the AI bubble alive. The less users understand how technology works, the easier it is to sell them unrealistic promises. The less leaders comprehend its scope, the more likely they are to make misguided decisions. And the less trained communicators are, the more space is given to narratives of miracles or fear. This knowledge gap is ultimately the foundation on which the illusion of automatic progress is built.

Reversing this situation requires a paradigm shift. Real digital literacy must go beyond technical instruction and become an integral education in technological thinking. This means teaching not only how to use technology, but how to interpret it, question it, and contextualize it. It requires integrating ethics, philosophy, and the social sciences into technological training, and vice versa. Only a society capable of understanding how and why the systems governing it work will be able to decide whether it truly wants them to do so. Because as long as knowledge remains in the hands of a few, artificial intelligence will continue to be an instrument of power, not progress.

8.2. Political decisions based on marketing rather than knowledge

One of the most characteristic features of the global expansion of artificial intelligence has been the speed with which governments have joined the technological race, often without fully understanding the scope or implications of their own decisions. Contemporary technology policy has become a spectacle of promises and headlines, where the priority is not knowledge but visibility. States compete to appear innovative rather than to be innovative, presenting national artificial intelligence plans, digital strategies and investment programs that, in most cases, lack a solid technical or scientific foundation. Artificial intelligence, rather than a tool for structural transformation, has become a product of political marketing.

In this dynamic, the discourse of innovation has prevailed over strategic reflection. Political leaders, aware of the symbolic value of technology in public opinion, use artificial intelligence as a form of modern legitimization. Including the term AI in speeches, budgets or national plans is equivalent to displaying modernity, progress and a forward looking vision, even when the actual content of those policies is superficial or redundant. Billion dollar digitalization programs are announced without evaluation mechanisms or concrete objectives; artificial intelligence observatories are created without researchers; chairs or centers of excellence are promoted that exist more in rhetoric than in practice. This phenomenon is repeated at all levels, local, national and international, and in virtually all regions of the world.

The European Union, for example, has developed multiple artificial intelligence strategies since 2018, focused on trust and ethics, but with a clear disconnect between declared principles and real implementation. In many cases, the funds allocated to innovation end up fragmented into pilot projects that generate almost no impact. The United States, for its part, followed a laissez faire policy until geopolitical pressure with China forced it to react through executive orders that combine regulation with national security rhetoric. In both cases, political language is filled with technical terms used more for their communicative effect than for their meaning, responsible AI, digital sovereignty, data economy. Discourse replaces knowledge.

The politicization of artificial intelligence is also evident in the way governments adopt regulatory decisions. Approval of the AI Act in Europe, for example, has been celebrated as a historic step in digital rights, but many experts have warned that its bureaucratic complexity and lack of flexibility could hinder innovation more than protect it. This contradiction reveals a recurring pattern, public policies on AI do not always stem from technical or scientific knowledge but from the desire to project political leadership. The result is legislation more oriented toward reassuring citizens and generating headlines than toward solving the structural challenges of the technology.

At the national level, many governments replicate this trend through digital theater strategies, campaigns announcing the incorporation of AI into education, healthcare or public administration without defining how, with what resources or under what criteria it will be implemented. In many cases, the projects remain in pilot phases or depend on private providers who control development and infrastructure. This creates technological dependence disguised as modernization. Governments present themselves as drivers of innovation, but in practice they merely finance and legitimize the interests of large technology corporations.

The lack of technical knowledge in political decision making worsens the problem. Most public officials lack training in data science, algorithms or digital governance. This makes them vulnerable to the influence of advisers, lobbies and consulting firms that act as interpreters of technological truth. These entities, with clear economic interests, write reports, set priorities and design national strategies in the name of scientific neutrality. In this way, decision making power shifts from parliaments to corporate boardrooms, and technology policy becomes an exercise in public relations.

The most dangerous consequence of these marketing driven decisions is the creation of unrealistic expectations among citizens. Governments promise a digital revolution that will solve unemployment, bureaucracy and inequality, but rarely explain the risks or limits of artificial intelligence. Automation is presented as synonymous with progress, without addressing its effects on work, privacy or social cohesion. This one sided narrative fuels technological hype and perpetuates the bubble, the citizen votes believing in a promised future, and the politician gains legitimacy through that same mirage.

In the long term, the consequences of this political superficiality are profound. Poorly planned investments lead to waste of public resources, improvised regulations slow down genuine innovation and social distrust in technology increases when promises are not fulfilled. In addition, the lack of coherence between policy and technical reality weakens the digital sovereignty of countries, making them dependent on large global platforms. Instead of building sustainable innovation ecosystems, showcases are built.

Solving this problem requires a transformation in the way technology policy is conceived. Artificial intelligence cannot be treated as a propaganda instrument or a campaign accessory. It requires interdisciplinary knowledge, dialogue among experts, institutions and citizens, and a long term vision that transcends electoral cycles. Without a deep understanding of its nature, AI will continue to be a decorative object in political discourse, a fashionable word useful for generating headlines but incapable of producing real progress. As long as politics continues to be guided by perception rather than knowledge, artificial intelligence will be administered with faith rather than reason.

8.3. Public procurement of systems without audits or metrics

One of the most concerning aspects of the uncontrolled expansion of artificial intelligence is the way governments and public administrations are acquiring systems and technological solutions without rigorous audits or verifiable metrics on performance, ethics, or social impact. This phenomenon, increasingly widespread, represents one of the main fractures in modern technological governance. In the name of efficiency, modernization, or digitalization, AI systems are being implemented in sensitive areas such as justice, healthcare, education, public security, or labor management without an adequate oversight framework, without transparency, and in many cases without truly understanding how they work. The result is a model of public AI adoption based more on blind trust and institutional marketing than on evidence and accountability.

The process usually begins with rushed procurement procedures driven by political pressure to keep up in the digital race. Instead of analyzing real needs or evaluating the relevance of proposed solutions, administrations buy technology for symbolic purposes: to demonstrate innovation, attract investment, or improve their image before public opinion. In many cases, these acquisitions are carried out through contracts with large corporations or tech startups that offer turnkey systems with promises of automation, cost savings, or advanced prediction. However, few of these contracts include clauses requiring algorithm audits, validation of results, or the measurement of impacts on fundamental rights. Once installed, these systems operate with almost total opacity.

The consequences of this lack of oversight are already visible. In the judicial sphere, several countries have implemented AI systems to support decisions on criminal risk or release eligibility, such as the controversial COMPAS system in the United States, without external audits or bias verification. In education, automated grading algorithms have shown discrimination based on origin or gender. In the public sector, AI programs designed to detect tax fraud or irregularities in subsidies have been denounced for privacy violations and for generating false positives that harm innocent citizens. Despite this, most of these projects remain operational without review, protected by contractual secrecy or by the technical complexity of the models.

The central problem lies in the absence of standardized metrics that allow objective evaluation of the performance and fairness of the acquired systems. While other areas of public management have financial, environmental, or quality audits, in artificial intelligence the assessment is left to the providers themselves. Governments buy promises, not verifiable results. In many cases, contracts include confidentiality clauses that prevent public access to the code, training data, or decision criteria of the system. This lack of transparency creates a vacuum of responsibility: when the technology fails, no one can clearly determine whether the error originated in the design, the use, or the political decision to implement it.

The absence of external audits is especially serious because artificial intelligence, by its nature, tends to amplify the biases and errors already present in the data. A model trained on incomplete or biased information can generate discriminatory decisions which, when applied in public environments, acquire institutional force. In this way, social prejudices become digitalized and automated. Without independent review mechanisms, governments run the risk of legitimizing injustice under the appearance of objectivity. Automation does not eliminate human error; it encodes it.

Furthermore, the problem is not limited to countries with low regulatory capacity. Even in advanced democracies, the acquisition of AI systems without rigorous audits is a common practice. In 2024, a report from the European Parliament revealed that more than 70 percent of public procurement processes related to artificial intelligence in the EU lacked formal mechanisms for ethical or technical impact assessment. In Latin America and Africa, the situation is even more alarming: many acquisitions occur through donations or agreements with foreign companies, creating structural technological dependency and a total lack of sovereignty over data.

The lack of standardized metrics also prevents comparison and continuous improvement. Each institution defines its own success indicators, often focused on economic efficiency such as reduced costs, faster processing, or the number of cases handled, without considering qualitative aspects such as fairness, transparency, or the protection of rights. This instrumental view of AI, where only the quantifiable is measured, reinforces the idea that technology is always good if it saves money, even if it degrades justice, privacy, or public trust.

This dynamic also has a harmful effect on the relationship between citizens and the state. When public decisions are automated without oversight, individuals lose the ability to understand and question the decisions that affect them. If an automated system denies a subsidy, a scholarship, or an authorization, citizens rarely know why or have an effective way to appeal. The lack of technical explainability becomes a lack of democratic explainability. Artificial intelligence, without audits, not only automates processes; it also automates the opacity of power.

To reverse this trend, it is urgent to incorporate independent audits and verifiable metrics at all levels of public procurement of AI systems. Administrations must demand algorithmic transparency, data traceability, and periodic review of results. They must also create public or mixed institutions with technical and ethical competence to supervise these technologies, just as there are courts of audit or anti-corruption agencies. Without these controls, governments will continue to feed a cycle of unreflective adoption that benefits large corporations, weakens technological sovereignty, and erodes public trust.

The lack of audits and metrics is not only a technical problem but also a political and moral one. It means that states are delegating their democratic responsibility to systems they do not understand, administering human decisions through automated mechanisms that no one supervises. And when accountability disappears, artificial intelligence ceases to be a tool serving the common good and becomes a modern form of unchecked power.

8.4. Fast-track training of “AI experts” with zero practical experience

One of the most obvious symptoms of the excessive boom in artificial intelligence is the proliferation of “AI experts” trained in a matter of weeks or even days, many of whom lack practical experience, deep technical understanding, or real knowledge of the field. This phenomenon, driven by market demand, educational marketing, and a social obsession with not being left behind, has generated an inflation of titles, certifications, and professional profiles that directly contribute to the inflation of the technology bubble. In just a few years, the term “AI expert” has lost its meaning, turning into a label of symbolic value rather than a credential of real competence.

The cause of this overproduction of superficial specialists lies in the combination of three factors: the urgency of the labor market, the opportunism of training institutions, and the psychological effect of hype. In response to the media explosion around artificial intelligence, thousands of universities, online academies, and training platforms have launched fast-track courses that promise to turn anyone into a professional in the field. Most of these programs focus on the use of pre-packaged tools or the handling of interfaces, without addressing the mathematical, statistical, or ethical foundations that underpin how AI works. People are taught to “use” models, not to understand them. The repetition of processes is promoted, not the capacity for analysis. And what is worse, certifications are granted that endorse skills that do not actually exist.

In this context, the labor market has been flooded with profiles that master the discourse but not the practice. Professionals who talk about neural networks, deep learning, or generative models without ever having implemented one from scratch. Consultants who advise companies on AI strategies without understanding how a model is trained, how its performance is evaluated, or what its maintenance entails. Instructors who teach artificial intelligence without ever having worked with real data. This trivialization of technical knowledge creates a self-reinforcing chain of incompetence: “experts” without experience train new “experts” who repeat the same conceptual and methodological mistakes, perpetuating the illusion of knowledge.

The problem does not lie only in the lack of technical depth, but also in the cultural distortion created by this phenomenon. Instead of fostering a scientific mindset based on rigor, experimentation, and doubt, a culture of immediacy has taken root, where learning is measured by speed rather than understanding. AI courses are advertised as fast-consumption products, promising professional success, promotions, and social recognition. Learning ceases to be a process of discovery and becomes a commercial transaction. Training turns into spectacle, and knowledge into merchandise.

This accelerated training has direct consequences for the quality of projects and trust in the sector. Companies that hire supposed specialists often discover that the results do not meet expectations. Projects fail not because of a lack of resources, but because of a lack of real knowledge. Models are trained with inadequate data, interpretations are made without rigor, and decisions are based on poorly informed intuition. Inexperience disguises itself as innovation, and improvisation is presented as creativity. This cycle of technical incompetence not only generates economic losses but also damages the reputation of the field of artificial intelligence itself.

The impact of this false professionalization also extends to public policy and media outreach. Many of the advisers, trainers, or communicators who currently define the AI narrative lack practical experience but have a great influence on public opinion. Their discourse, based on simplifications and attractive metaphors, fuels sensationalism and reinforces the bubble. By turning AI into a fashionable topic rather than a field of knowledge, they distort the collective perception of what it really means to develop, implement, and maintain artificial intelligence systems. In this way, technical ignorance is disguised as intellectual leadership, and complexity is replaced by empty phrases such as “AI will change the world” or “we need to use AI in everything.”

The acceleration of training has also created a dangerous paradox: as the number of “experts” increases, true technical competence decreases. The most advanced technology companies face a real shortage of professionals capable of understanding the fundamentals of machine learning, model architectures, or large-scale data management. The abundance of certificates has not solved the lack of knowledge; it has hidden it. This gap between the apparent supply of talent and the real capacity for execution is one of the structural factors that feeds the fragility of today’s AI ecosystem.

Ultimately, the phenomenon reflects a deeper crisis: the loss of the value of knowledge in favor of the value of appearance. Education, instead of being a process of building judgment and experience, has become a personal marketing strategy. Being perceived as an expert matters more than actually being one. And in an environment dominated by social networks and the attention economy, this dynamic multiplies. The same inexperienced “experts” become tech influencers, offering courses, writing articles, or giving talks on topics they barely understand, thus contributing to the creation of a self-referential ecosystem where visibility replaces competence.

Reversing this situation requires rebuilding the notion of training in artificial intelligence from the ground up. It is necessary to establish international standards that distinguish between basic, professional, and specialized training, and that require real practical experience before granting accreditations. Universities and training centers must recover scientific rigor and promote interdisciplinarity, integrating ethics, philosophy, and sociology alongside technical knowledge. And companies, instead of hiring based on titles or buzzwords, must evaluate analytical capacity, real project experience, and an understanding of the limits of the technology.

The bubble of false AI experts is an exact reflection of the AI bubble itself: both are sustained by the illusion of knowledge. As long as the market rewards speed over depth and appearance over competence, artificial intelligence will continue to grow on fragile foundations. Because in the end, true intelligence, whether artificial or human, is not measured by the ability to repeat what has been learned, but by the ability to understand what is not yet known.

8.5. The danger of institutionalized ignorance and the delegation of critical decisions to opaque systems

One of the most serious and least discussed risks surrounding the expansion of artificial intelligence is that of institutionalized ignorance, understood as the widespread adoption of technological systems whose complexity exceeds the understanding of those who administer or oversee them. In this scenario, institutions such as governments, corporations, financial entities, international organizations, or public administrations end up delegating critical decisions to algorithms they do not understand, trusting in their apparent neutrality and precision. This delegation is not only technical but also cultural and political: it represents a progressive renunciation of human responsibility in the name of efficiency, objectivity, and the supposed infallibility of the machine.

Institutionalized ignorance emerges from a mismatch between the speed of technological development and the slowness of institutional adaptation. Organizations pressured by competition, markets, or public opinion adopt artificial intelligence solutions without having the internal capacity to evaluate, audit, or understand them. Those responsible for implementation often lack training in data science, statistics, or technological ethics, and depend on external providers who control the knowledge and processes. As a result, institutions become passive users of systems that operate as black boxes, where decisions are the outcome of calculations impossible to trace or interpret. Power shifts from human analysis to automation, and with it the capacity to deliberate, question, and correct.

This phenomenon has direct implications for democratic governance and organizational integrity. When an institution delegates sensitive decisions such as granting loans, allocating resources, selecting personnel, approving medical treatments, or monitoring a population to opaque systems, the chain of responsibility erodes. No one can explain or assume the consequences of an error because the process is considered automatic. What used to be a debatable decision is now presented as a technical result, supposedly free of subjectivity. This illusion of objectivity turns the algorithm into the supreme arbiter of reality, yet one that is accountable to no one.

Blind delegation to opaque systems is dangerous not only because of what they decide but also because of what they make invisible. Automation tends to naturalize decisions, presenting them as inevitable. When an algorithm determines hospital priority or eligibility for a subsidy, the citizen can no longer appeal to reason or dialogue but only to a mathematical output. This weakens democratic mechanisms of participation and transparency, replacing public debate with a technocratic administration of power. Authority no longer comes from knowledge or deliberation but resides in the machine. What is lost in this process is not only institutional control but also the sense of humanity in decision making.

The problem is aggravated by the structural lack of transparency of many AI systems. Developer companies, shielded by intellectual property, refuse to reveal the data, models, or training processes behind their algorithms. Governments and corporations thus use tools they cannot audit or understand, relying on the goodwill of the provider. This model of technological dependency creates an unprecedented power asymmetry: public and private institutions become mere executors of decisions designed by external entities that control knowledge and data. Institutional sovereignty dissolves, and with it the capacity for autonomous decision making.

Institutionalized ignorance also has an ethical dimension. Delegating critical decisions to opaque systems eliminates moral responsibility. If a company dismisses employees based on an algorithmic recommendation, or if a judicial system assigns a sentence using a predictive model, who is responsible for the outcome? The absence of technical understanding facilitates the evasion of responsibility, shifting blame to the machine, the model, or the provider. Automation becomes an excuse. Under the argument of efficiency, institutions dehumanize their decisions and dilute accountability, replacing justice and prudence with speed and supposed accuracy.

This phenomenon is not only a governance risk but also a threat to social stability. Algorithmic opacity generates public distrust, especially when errors or biases become visible. Cases of algorithmic discrimination in hiring systems, facial recognition, or the granting of social benefits have shown that automation without understanding can reproduce and amplify existing inequalities. Yet instead of questioning the system, many institutions choose to defend it, citing technical ignorance or lack of alternatives. Ignorance becomes institutionalized when it ceases to be a limitation and becomes an implicit policy, that of blindly trusting what is not understood.

The deepest risk of this dynamic is the loss of collective intellectual autonomy. A society that delegates to opaque systems not only loses control over its processes but also over its capacity to learn. Institutions stop analyzing, investigating, or questioning, confining themselves to managing outputs. Artificial intelligence, instead of expanding human intelligence, ends up replacing it. At that point, knowledge ceases to be a social value and becomes a subcontracted service.

To prevent institutionalized ignorance from taking root, it is essential to reform the relationship between institutions and technology. This requires three pillars: transparency, training, and responsibility. Systems used by governments or corporations must be auditable and explainable; institutions must invest in the technical and ethical training of their decision makers; and automated decisions must always have a clearly identified human responsible. Artificial intelligence can be an extraordinary support tool but should never become a substitute for human judgment.

Because when a society accepts that its institutions no longer understand the tools that govern their decisions, it has crossed a dangerous line, the line where knowledge is abandoned as the basis of power. At that moment, intelligence ceases to be a human attribute and becomes a statistical illusion. And civilization, instead of advancing toward an era of augmented wisdom, enters a new form of darkness, one in which data govern but no one understands what they mean.

9. Possible future scenarios

The outcome of the current artificial intelligence bubble will depend on how governments, companies, and society manage the tensions between innovation, regulation, sustainability, and technological realism. At this point, the question is no longer whether the bubble exists, but how and when it might transform or collapse. Economic and technological history shows that every speculative frenzy reaches a tipping point at which expectations can no longer sustain perceived value, and collective confidence fractures. In the case of AI, that moment could be especially complex, because what is at stake is not only the invested capital but also the digital infrastructure that supports much of the economy, education, science, and public administration.

Exploring the possible scenarios is not an exercise in prediction, but in responsible analysis. Each outcome, from an abrupt collapse to a controlled slowdown or a market reconversion, would have different implications for the global economy, social stability, and the evolution of technological thinking. Understanding these potential pathways allows not only for preparation ahead of the crisis but also for recognizing an opportunity: redirecting the development of artificial intelligence toward a more sustainable, ethical, and human-centered model.

9.1. Scenario A: the bubble bursts and triggers a crisis of confidence and capital similar to 2001

The first possible scenario is the most drastic: the abrupt bursting of the artificial intelligence bubble, triggering a crisis of confidence and capital comparable to the collapse of the dot-com bubble in 2001. In this scenario, the market finally confronts the economic and technical reality of the sector: thousands of startups without a viable business model, overvalued companies unable to monetize their AI-based products, and investment funds discovering that promises of exponential growth were unsustainable. The excess of expectations, combined with the exhaustion of speculative capital and the saturation of a market with no real differentiation, acts as the trigger.

The process could begin with the fall of one or several major companies considered pillars of the ecosystem, whose financial results fail to meet the growth projections the market expected. As happened with Pets.com, Webvan, or eToys in 2001, the collapse of certain key players would serve as a catalyst for widespread distrust. Investors, faced with the lack of tangible profits, would begin withdrawing their support from projects inflated by hype, causing a chain reaction of shutdowns, massive layoffs, and devaluation of technological assets. This domino effect would impact both emerging startups and established giants that rely on the AI narrative to sustain their market value.

The warning signs of such an explosion are already visible. The excessive growth in valuations of generative AI companies, the competition for funding without differentiated products, and reports of multimillion-dollar losses in leading firms are symptoms of an economy inflated by belief rather than results. A particularly illustrative case would be companies offering cloud-based language models or image generation: services that are extremely costly to maintain, with narrow profit margins and a volatile user base. If revenue fails to cover the costs of computing, energy, and data licensing, even the most popular projects could become financially unsustainable.

The economic impact of a collapse of this magnitude would be immediate. Thousands of jobs would disappear, especially in tech-driven startups, consulting firms, and corporate innovation departments. Venture capital funds, which in recent years have poured billions into AI projects without clear returns, would drastically reduce their exposure, also affecting adjacent sectors such as cloud computing, cybersecurity, and data analytics. Large technology corporations, although more resilient, would suffer significant stock market declines and cuts in their research and development divisions. The psychological shock to the market would resemble that of 2001: the belief that “this time was different” would be proven false.

In the media and society, the collapse would trigger a crisis of confidence in artificial intelligence overall. Headlines would shift from euphoria to disappointment, and public opinion would begin to view AI not as an inevitable revolution but as yet another bubble in the history of technological capitalism. The narrative of automatic progress would crumble, giving way to widespread skepticism toward new technological promises. Distrust would not only affect investors but also consumers, who would become wary of AI applications in everyday life, especially in areas such as education, healthcare, or employment.

The impact of this scenario would not be only economic but also political. Governments, pressured by social unrest and job losses, could tighten technological regulation or even freeze investment programs in AI, slowing innovation on a global scale. At the same time, a populist discourse could emerge against big tech corporations, blaming them for inflating the bubble and speculating with public expectations. The aftermath of the collapse could lead to a period of contraction and distrust similar to the post-dot-com era: a painful but necessary adjustment to purge the ecosystem and separate genuine innovation from marketing disguised as science.

However, as occurred after 2001, a crisis of this magnitude could also act as a catalyst for a new stage of technological maturity. After the dot-com collapse, companies emerged that understood the limitations of the previous model and developed sustainable structures, such as Amazon or Google. In the same way, the bursting of the AI bubble could cleanse the market of speculative projects and pave the way for a second generation of companies more focused on real utility, sustainability, and ethics. But before that can happen, this scenario would involve a deep shock to the digital economy and a massive loss of capital and confidence.

In historical terms, this scenario represents the classic cycle of euphoria, overinvestment, and collapse that accompanies every major technological transformation. The difference is that, in the case of artificial intelligence, the consequences would not be limited to the financial sphere: they would affect the cognitive infrastructure of modern society, the relationship between humans and technology, and the very credibility of the scientific narrative. The bursting of the AI bubble would not be only an economic crisis but also a symbolic one: the end of the illusion that intelligence, whether natural or artificial, can replace human judgment without consequences.

9.2. Scenario B: the bubble deflates slowly, with market consolidation and the survival of the strongest projects

The second possible scenario is less abrupt but equally significant: the gradual deflation of the artificial intelligence bubble, a slow adjustment process in which the initial enthusiasm gives way to economic, technical, and social realism. In this scenario, there is no sudden collapse but a sustained correction over several years, during which the market, investors, and institutions gradually acknowledge the limits of the current model. The euphoria fades without bursting, and the AI ecosystem enters a maturation phase where only projects with solid foundations, long-term vision, and real value manage to survive.

The deflation of the bubble would follow a well-known pattern in technological history: the natural cleansing of an oversaturated market. As investments slow down and speculative capital withdraws, companies relying solely on hype begin to disappear or are absorbed by larger ones. Startups offering redundant or undifferentiated solutions find themselves unable to sustain operations once easy funding vanishes. Meanwhile, established corporations restructure their AI divisions, prioritizing profitability and strategic integration over massive experimentation. The result is a process of natural selection: the initiatives that survive are those providing tangible value, solving concrete problems, and sustaining themselves through viable business models.

This scenario would entail a profound shift in industry mindset. The narrative of “AI as a universal solution” would give way to a more pragmatic approach focused on efficiency, safety, and sustainability. Companies would stop investing out of trend and begin demanding measurable returns. Artificial intelligence projects would shift from media-driven experiments to operational infrastructures integrated into sectors such as logistics, energy, agriculture, healthcare, or education. The market would leave behind the phase of promises and move into the era of utility.

On the financial level, the deflation process would translate into a moderate but prolonged contraction of valuations. Large tech companies would experience declines in stock prices, though not collapses; venture capital funds would adopt more selective strategies, prioritizing projects with real impact; and mergers and acquisitions would increase as survival mechanisms. This less speculative environment would favor the concentration of power in a few dominant players, mainly large corporations controlling infrastructure, data, and computation, while smaller companies without financial backing would be absorbed or disappear. This would lead to a market consolidation similar to what followed the dot-com bubble, when a few companies, such as Google, Amazon, or eBay, emerged strengthened from the ruins of widespread enthusiasm.

On the technical front, deflation could have a positive effect by allowing research to slow down and regain lost rigor. Instead of chasing headlines or new model versions every few months, the scientific and business communities could refocus on solving structural problems such as energy efficiency, semantic understanding, or algorithmic transparency. Universities and research centers, less pressured by market urgency, could recover their role as spaces for genuine and critical innovation. In this sense, the end of euphoria could mark the beginning of a new phase of intellectual maturity for artificial intelligence.

In the social sphere, a gradual deflation process would also mitigate the psychological impact of lost confidence. Instead of an abrupt collapse in expectations, there would be a progressive adaptation: society would learn to view AI with greater critical distance, recognizing its real capabilities without attributing it messianic qualities. Users, businesses, and institutions would understand that artificial intelligence is not a magical force but an imperfect, powerful, and useful tool within certain limits. This cognitive adjustment could give rise to a more balanced relationship between humans and technology, less driven by fascination and more by responsibility.

However, this scenario also carries risks. Market consolidation could reinforce technological monopolies, increasing global dependence on a handful of corporations capable of sustaining the development and operational costs of advanced models. This would widen power gaps between regions and sectors, pushing grassroots innovation into the background. Additionally, the decline in speculative investment could reduce the diversity of ideas and the dynamism of the ecosystem, sidelining alternative projects that, while not immediately profitable, could have long-term strategic or social value.

From a political and regulatory standpoint, deflation would allow governments to act with greater calm. With the pressure of hype reduced, artificial intelligence policies could shift toward education, ethical regulation, and investment in sustainable digital infrastructures. Regulatory frameworks, such as the European AI Act, could be applied with fewer tensions and greater clarity. Likewise, the fall in excessive enthusiasm would create space for a more rational debate on the limits of automation, job protection, and the preservation of human thought in the face of technological delegation.

This scenario describes a process of maturity rather than a crisis. The bubble does not burst; it slowly deflates, leaving behind a more stable industry, deeper knowledge, and a more aware society. The price of this adjustment would be the abandonment of the illusion of infinite growth, but the benefit would be a more sustainable and lucid relationship with technology. Artificial intelligence would cease to be a symbol of unfulfilled promises and become a critical infrastructure of the modern world: less spectacular but more useful, less idolized but better understood.

9.3. Scenario C: governments intervene and stabilize the system through regulation and education, slowing down the collapse

The third possible scenario proposes an intermediate path between abrupt collapse and natural deflation: a partial stabilization of the system driven by active government intervention. In this scenario, political authorities, aware of the systemic risk represented by the artificial intelligence bubble, act to contain it through a combination of regulation, market oversight, strategic investment, and above all, public education. The correction is not avoided, but its impact is softened. The collapse does not occur violently; instead, it becomes a controlled slowdown that allows the technological ecosystem to be redirected toward a more sustainable, transparent, and human model.

The trigger for this intervention is usually the accumulation of warning signs: mass layoffs, technological fraud, scandals related to the misuse of data, failures in critical automated systems, or growing social unrest linked to automation. At that point, governments understand that artificial intelligence is not only an economic issue, but also a matter of social and strategic stability. International institutions such as the European Union, the OECD, or the United Nations begin coordinating regulatory efforts, while national governments establish laws requiring technology companies to comply with standards of transparency, traceability, and accountability. This regulatory process, far from stifling innovation, aims to rebalance it by setting clear limits that reduce the risk of abuse, speculation, and concentration of power.

In economic terms, governments could choose to intervene directly in capital flows. Faced with shrinking private investment, public funds and innovation programs would take on a compensatory role, financing projects with long-term social and scientific value that go beyond immediate profitability. In this way, artificial intelligence would be steered toward strategic sectors such as health, education, energy sustainability, and public administration. State intervention, combined with tax incentives and conditional subsidies, would help maintain activity in the sector while filtering out speculative excess. The goal would not be to reignite the hype, but to channel technology toward real and verifiable objectives.

Regulation would be the other central pillar of this scenario. Frameworks such as the European AI Act, the United States Executive Order on Artificial Intelligence, or China’s algorithmic oversight policies would be strengthened and harmonized to avoid an uncontrolled race. Mandatory mechanisms for auditing, ethical assessment, and model certification would be established before large-scale deployment. In addition, companies would be required to justify the environmental and social impact of their systems, assuming responsibility for their effects. Although such measures would slow the pace of innovation, they would also reduce the risk of collapse by creating a more predictable and reliable environment. The immediate consequence would be a less volatile market and a reduction in speculative appeal, but one that is more balanced in the long term.

The third component of this scenario — and perhaps the most transformative — would be education. Governments, aware that the lack of digital literacy and critical thinking has fueled both the bubble and the sensationalism surrounding it, would launch large-scale educational programs to train citizens and professionals in the conscious and responsible use of technology. Artificial intelligence would no longer be taught as a magical tool, but understood as a social and ethical infrastructure. Schools, universities, and vocational training centers would incorporate subjects on algorithms, data, privacy, and digital ethics, while public campaigns would promote a realistic understanding of the role of AI in everyday life. In this way, knowledge would become a preventive tool against technological manipulation.

At the international level, government intervention could lead to a new geopolitical balance. States that manage to combine innovation with effective regulation and public education would position themselves as leaders of the next phase of technological development. Europe, with its ethical and regulatory approach, could become the normative reference; the United States, the center of business and scientific development; and Asia, the engine of mass production and adoption. This diversity of approaches could generate tensions but also a more stable and plural system in which technological power does not depend exclusively on speculation or private monopolies.

However, this scenario is not free from risks. Excessive intervention could lead to bureaucratization and slowness, reducing competitiveness and discouraging innovation. Likewise, if regulation is used as a political or ideological tool, it could limit research freedom or encourage technological censorship. The key would be to find a balance between control and creativity, between security and progress. Too many restrictions could smother the sector’s dynamism, while the absence of them would lead to the chaos this scenario seeks to avoid.

In this scenario, the future of artificial intelligence would depend less on market forces and more on the political and cultural maturity of societies. Regulation, education, and ethics would replace speculation as the primary drivers. Governments would become central actors in the new technological era, not as censors, but as guarantors of balance between innovation and responsibility. Progress would no longer be measured in terms of speed or spectacle, but in sustainability, social impact, and collective benefit.

The outcome of this scenario would be a technological ecosystem that is slower but more stable; less lucrative but more human. The bubble would not burst or dissolve completely, but would transform into a consolidation phase guided by knowledge and prudence. Artificial intelligence, far from disappearing, would be redefined as a global public infrastructure governed by oversight, education, and purpose. It would be, in a way, the triumph of intelligence over faith.

9.4. Scenario D: the cycle continues and humanity does not learn, transferring the same dynamic to another emerging technology (for example, biotechnology or quantum technology)

The fourth scenario is both the most pessimistic and the most likely: the cycle continues, and humanity, far from learning from experience, repeats the pattern. The artificial intelligence bubble does not burst or deflate in a transformative way; it simply loses prominence and is replaced by a new technological promise. When media attention and capital become exhausted, collective enthusiasm shifts toward the next frontier: biotechnology, quantum computing, neurotechnology, or any other emerging field that promises an even deeper “revolution”. Just as AI replaced blockchain, and blockchain replaced big data or the internet of things, the focus will shift toward a new savior idea. The bubble does not die: it mutates.

This scenario is based on an undeniable historical reality. Since the Industrial Revolution, every wave of innovation has been accompanied by an identical cycle: discovery, enthusiasm, massive investment, speculation, saturation, crisis, and adjustment. The difference is that in the digital age these cycles have accelerated. The speed with which the market and public opinion consume technological narratives has reduced the time between euphoria and disappointment. What once lasted decades now occurs in a matter of years. And, as with any addiction, society seeks increasingly intense stimuli to maintain the illusion of uninterrupted progress. If artificial intelligence loses its power of fascination, the economic and media system will quickly find a new promise to worship.

In this context, technologies such as biotechnology or quantum computing emerge as the next natural candidates. Biotechnology, driven by advances in genetic editing, neuroscience, and brain-machine interfaces, promises a literal fusion between the biological and the digital, reviving the myth of the enhanced human being. Quantum computing, on the other hand, offers a narrative of mathematical transcendence: the possibility of breaking the limits of classical physics and accessing a new, almost metaphysical level of processing. Both disciplines share the ingredients of the next major hype: technical complexity inaccessible to the general public, potentially infinite applications, and a symbolic charge of salvation —the cure for cancer, the control of aging, the resolution of climate change, the total simulation of reality.

The problem does not lie in the technologies themselves, but in the way societies interpret them. When faith in progress prevails over critical thinking, every new innovation becomes a new dogma. Artificial intelligence has demonstrated the extent to which mechanisms of speculation, marketing, and manipulation can distort scientific knowledge. If humanity does not correct this dynamic, it will transfer the same mistakes to any other emerging field. Messianic narratives will reappear under different names: “conscious biology”, “divine computing”, or “genetic singularity”. Industry and media will once again promise a perfect future, and the public will once again believe.

In the economic sphere, the repeated cycle would have predictable consequences. The same investment funds that inflated the AI bubble would redirect their capital toward the new frontier, inflating valuations and generating a new speculative fever. Startups, universities, and governments would compete to present themselves as pioneers, repeating the pattern of overinvestment without scientific validation or sustainable models. The logic of immediate profitability would once again prevail over that of deep understanding. When the initial enthusiasm fades, history will be written again with the same words: “failure”, “market correction”, “crisis of confidence”.

The repetition of the cycle would also have a psychological and cultural component. Contemporary human beings live trapped in a paradox: they fear stagnation as much as uncertainty. That is why they constantly need to believe in the existence of a new frontier. Technology has become the modern substitute for religious faith, offering a horizon of collective and personal salvation. Each technological bubble represents not only an economic cycle but also an emotional cycle: promise, hope, disappointment, and the search for a new idol. In this sense, the inability to learn from mistakes is not only economic but existential. Humanity clings to the illusion of unlimited progress because it fears facing the void left by its absence.

On the political level, this scenario would perpetuate the structural dependence of states on technological corporations. Governments, instead of acting as conscious regulators, would follow the same pattern of late reaction: first celebrating innovation, then investing in it without control, and finally attempting to regulate it when the damage is already visible. Public policies would adapt to the media cycle, not to scientific evidence. Thus, technological governance would continue to be reactive rather than preventive, reproducing the same dynamic of improvisation and misinformation that has characterized the era of AI.

The greatest risk of this scenario is not economic but civilizational. If human beings continue to delegate their sense of progress to technological forces they do not understand, they will end up trapped in a loop of intellectual dependence. Each bubble will not only destroy capital but also erode trust in science, in education, and in the human capacity for reason to guide the course of knowledge. Progress will become a continuous spectacle, a sequence of promises that replace previous ones before being verified. The ultimate consequence of this dynamic would be the trivialization of the future: a state of perpetual technological euphoria that, paradoxically, prevents genuine advancement.

However, even within this scenario of repetition, there is a possibility of redemption. Collective memory, although fragmented, always leaves traces. Each cycle leaves behind a critical minority of voices who learn, document, and reflect. These are the voices that, over time, build the foundations of authentic knowledge that allows civilizations to evolve. If humanity cannot avoid repeating the cycle, it can at least aspire to shorten it: to reduce the time between euphoria and awareness. Ultimately, learning does not consist in never stumbling, but in getting up sooner.

This scenario therefore describes a humanity advancing in a spiral: always revolving around the same illusions, but each time with a slightly deeper understanding. And even if the cycle continues, each repetition leaves a trace, a warning, and an opportunity. Because even when history repeats itself, knowledge —if preserved— can ensure that in the next bubble the blow is less devastating and lucidity arrives a little earlier.

10. My humble opinion, as an expert, as a user, and as a member of this society we live in

The phenomenon of artificial intelligence is not, in essence, the problem. The real issue lies in what human beings have projected onto it: their fears, their ambitions, and their need for control. AI is nothing more than an amplified reflection of the human mind, a mirror returning the image of our deepest anxieties. In its development, we have pursued perfection, absolute efficiency, the elimination of error, omniscience. Deep down, we have not tried to create intelligence, but rather to build a substitute for our uncertainty. We have burdened it with the responsibility of thinking for us, of deciding better than us, of freeing us from the limits that make us human. We created it to take the weight of responsibility off our shoulders. But in that quest for control, what we have truly revealed is our inability to tolerate the imperfection, the doubt, and the fragility that are part of the human condition.

Artificial intelligence has become the stage on which humanity rehearses its eternal struggle between reason and desire. Every technical advance has been met with a mixture of fascination and fear, hope and dependency. The paradox is that while we try to build an external intelligence that solves everything, we are allowing our own to deteriorate. The AI bubble is not only economic or technological; it is intellectual. It is the result of a deeper crisis, that of human intelligence and the loss of critical thinking. In a world saturated with information, knowledge has been replaced by immediacy; reflection, by productivity; and understanding, by the appearance of understanding. We have delegated thinking to algorithms with the same ease with which we once delegated it to institutions or ideologies. And in doing so, we have confused knowing with accessing data, and thinking with processing information. So many years to evolve, and so little to lose it. Just the blink of an eye in the vastness of the eternal universe. Hypocrisy, ignorance, and modern illiteracy, what I call mental illiteracy, have turned millions of people into passive spectators of their own decline, unable to distinguish between what is real and what is fabricated, between what is profound and what is superficial, between what should matter and what merely entertains.

The crisis of artificial intelligence is, in reality, the visible symptom of a much older disease: the dehumanization of thought. When technology ceases to be a tool and becomes an end in itself, progress becomes empty of moral content. We have grown accustomed to measuring progress by the speed of change and not by its meaning, acting without common sense. We speak of innovation without asking what for, of automation without asking at what cost, and of intelligence without asking what it means to be intelligent. I have dealt with the meaning of the term intelligent since I was only seven years old, because I have an IQ above 140, and I can assure you that neither humans truly know what intelligence means, nor are they prepared for true intelligence. Human beings have always been deeply confused, and in that confusion, humanity has lost part of its inner compass. AI has not stolen our ability to think; we have abandoned it voluntarily, seduced by the comfort of a world that no longer demands intellectual effort or individual responsibility.

But every crisis, even a technological one, contains the possibility of renewal. The burst or deflation of the artificial intelligence bubble can become a historic opportunity to rebalance our relationship with technology and with ourselves. This turning point forces us to ask what kind of intelligence we want to cultivate: the one that calculates or the one that understands; the one that predicts or the one that interprets; the one that imitates or the one that creates. Perhaps the most important lesson AI can offer us is to remind us what it means to think from a human standpoint: with awareness, with empathy, with ethical intention.

Real progress does not lie in developing smarter machines, but in building a wiser society. That implies recovering the ability to question, to doubt, to reflect before automating. It means understanding that technology should not replace our intelligence, but expand it; should not dominate our will, but serve it. AI can be an extraordinary tool, but only if the purpose guiding it is deeply human. Otherwise, we will find ourselves in front of a mirror that becomes increasingly precise, yet increasingly empty.

The artificial intelligence bubble, with all its euphoria and excesses, can be seen as a generational mirror. It reflects the exhaustion of a model of thought that confuses knowledge with power and progress with accumulation. However, it can also be the starting point of a new era of technological maturity, one in which human beings stop asking what they can do with artificial intelligence and begin asking what they should do with it. The true challenge is not to perfect machines, but to reclaim the meaning of human thought. Because if this bubble teaches us anything, it is not what AI can achieve, but what humanity has stopped daring to do: to think deeply, act consciously, and progress with purpose. We are losing our essence, what makes us human. We have gone from playing outside to endless scrolling.

Sources:

1. “Global stock markets fall sharply over AI bubble fears” (The Guardian, 5 Nov 2025)
   https://www.theguardian.com/business/2025/nov/05/global-stock-markets-fall-sharply-over-ai-bubble-fears
2. “AI can be both a bubble and a breakthrough” (Reuters, 6 Nov 2025)
   https://www.reuters.com/markets/ai-can-be-both-bubble-breakthrough-2025-11-06/
3. “Is the A.I. Boom Turning Into an A.I. Bubble?” (The New Yorker)
   https://www.newyorker.com/news/the-financial-page/is-the-ai-boom-turning-into-an-ai-bubble
4. “Tech guru Erik Gordon says investors will ‘suffer’ far more from the AI boom than the dot-com crash” (Business Insider, August 2025)
   https://www.businessinsider.com/erik-gordon-ai-stocks-dot-com-bubble-crash-tech-market-2025-8
5. “Are We At Peak AI Bubble And The Cusp Of ‘AI Moment’?” (Forbes, March 2025)
   https://www.forbes.com/sites/gilpress/2025/03/30/are-we-at-peak-ai-bubble-and-the-cusp-of-ai-moment/
6. “Is the AI bubble about to burst – and send the stock market into freefall” (The Guardian, August 2025)
   https://www.theguardian.com/technology/2025/aug/23/is-the-ai-bubble-about-to-burst-and-send-the-stock-market-into-freefall
7. “What we mean when we talk about an artificial intelligence ‘bubble’” (World Economic Forum, October 2025)
   https://www.weforum.org/stories/2025/10/artificial-intelligence-bubble-dot-com-tulip-mania/
8. “The AI Bubble Is Leaking: Prepare For A Major Market Reversal” (Seeking Alpha)
   https://seekingalpha.com/article/4816134-the-ai-bubble-is-leaking-prepare-for-a-major-market-reversal
9. “Wall Street Is Worried About an AI Bubble—Here’s the Sector Where Stock-Prices Really Stand Out” (Investopedia, very recent)
   https://www.investopedia.com/wall-street-is-worried-about-an-ai-bubble-the-sector-where-stock-prices-really-stand-out-11835209
10. “Is a bubble forming as AI investments drive economic growth?” (PBS)
    https://www.pbs.org/newshour/show/is-a-bubble-forming-as-ai-investments-drive-economic-growth