AI Enters the 'Post-Hype Era': A Structural Shift from Technological Revelry to Global Governance

AI Enters the 'Post-Hype Era': A Structural Turn from Technological Revelry to Global Governance

In this week of August 2026, a series of seemingly unrelated news items together sketch an often-overlooked turning point in the age of artificial intelligence. Secondhand booksellers in the UK and Ireland have discovered a large number of "strange" bulk orders—old books being centrally acquired by unidentified buyers, with subsequent investigations pointing to data-collection activities related to AI companies. Meanwhile, Scottish police publicly warned that proposed data centers may face "serious security threats" and are expected to encounter "substantial public opposition."

These scattered incidents appear, on the surface, to be ordinary ripples in the technology cycle, but they actually reflect a more profound structural change: AI is falling from a purely technological revelry into a "post-hype era" marked by fierce collisions with the real world. At this stage, the deep-seated contradictions in data, energy, law, employment, culture, and geopolitics can no longer be masked by Silicon Valley's optimistic narrative.

I. The Battle for Data: AI's "Enclosure Movement"

The anomalous orders in the secondhand book market reveal a long-ignored reality: AI companies' demand for data has entered an almost pathological phase of expansion. As publicly accessible text on the internet approaches exhaustion, AI companies are pushing deeper into the physical world—books, journals, and even all printed matter. Anthropic is reported to have spent millions of dollars purchasing books for scanning, in what is called "data acquisition." This is not an isolated case, but a microcosm of a new "enclosure movement"—only this time, the object of contention has shifted from land to humanity's collective intellectual heritage.

This raises a critical question: should the entirety of human civilization's knowledge be incorporated by a few technology giants? Once books are scanned into training datasets, they are no longer pure cultural inheritance but become the underlying assets of AI models. Copyright boundaries, authors' rights, and the ownership of public knowledge have become unprecedentedly ambiguous in the new data scramble. Global regulators are not yet ready to answer these difficult questions, while technology companies are already several strides ahead.

II. The "NIMBY Dilemma" of Infrastructure

Scottish police's concerns about data center security reveal another face of the AI revolution. AI's computing demands are driving a frantic expansion of data centers worldwide, but wherever these facilities go, they often clash fiercely with local communities. The proposed data center near Edinburgh has been deemed likely to draw "substantial public opposition." This is not an isolated phenomenon—from Europe to Asia, from the United States to Latin America, issues of water, electricity, land, and environmental impact surrounding data centers are becoming a new political battlefield.

What is even thornier is the security dimension. The police explicitly regard data centers as potential attack targets, meaning that AI infrastructure is no longer merely a commercial asset but critical infrastructure linked to the lifeline of nations. Once data centers become objects of terrorism or retaliatory attacks between states, the information nervous system of the entire society could collapse. Such vulnerability has never before been presented on such a scale by any previous generation of general-purpose technology.## III. The Myth of AI Democratization and the Reality of Power

Against a backdrop of growing public anxiety, Mark Zuckerberg published a lengthy essay of roughly six thousand words, declaring that “superintelligent AI” should be “for everyone,” while simultaneously releasing Meta’s open-weight model. On the surface, this is a blessing for the democratization of AI; but in the eyes of critics, it is precisely another repackaging of an oligarchic agenda.

The narrative that “everyone owns AI” deliberately obscures a core question: Who truly controls the foundation models? Who profits from their deployment? Who can constrain the boundaries of use? Open-weight models do lower the barrier to entry, but they do not change the underlying power structure. The computing power, data, energy, and talent needed to train AI remain highly concentrated in a very small number of tech giants. In an article, Guardian commentator Raffi Krikorian pointed out that the key is no longer “whether AI belongs to everyone,” but “who actually owns it.”

The deeper contradiction is laid bare in market logic. Commentators Bruce Schneier and Nathan Sanders have even suggested that if OpenAI and Anthropic cannot be accepted by the market, the United States should nationalize them. This seemingly extreme argument reflects a new reality: the core assets of AI may inherently carry a public character, and the current corporate system cannot price their externalities. When a technology simultaneously possesses strategic, security, and social attributes, whether the market can effectively allocate resources has become a global governance challenge.

IV. The “Unexpected” Employment Crisis

For years, the mainstream narrative kept exaggerating that AI would destroy vast numbers of jobs. By 2026, however, reality presents a different picture—a Guardian column headline bluntly asks: “AI was supposed to destroy jobs, so where is the ‘carnage’?” The unemployment rate has not seen the predicted cliff-edge collapse, but this by no means means the job market is safe and sound.

What we are witnessing is a hidden restructuring of the employment structure. Unemployed young people are being sent into AI training camps—not so much to prepare for future new jobs as to passively adapt the labor market to technological shocks. Jobs have not disappeared, but the form, compensation, and stability of work are undergoing deep changes. AI is not eliminating large numbers of jobs at once, but slowly reshaping task allocation, performance evaluation, and employment relations. This chronic erosion is more difficult to address than sudden “technological unemployment,” because it cannot be quickly reversed by a single piece of legislation, but can only be absorbed through a long-term, systematic reshaping of the social security system.

V. The Legal and Moral Vacuum

When a teenager in Massachusetts was charged with using ChatGPT to kill his mother and brother, the question of AI responsibility fell from philosophical speculation onto the sharp blade of criminal courtroom reality. At the same time, Australia reported the first automated hacking incident by an AI agent, and experts immediately pointed out: the AI agent bears no legal responsibility for any harm it causes—responsibility ultimately falls on the deployer, and even on the developer.This exposes a troubling legal vacuum. Our legal system rests on the cornerstone of "human actors," yet the autonomy and unpredictability of AI systems are shaking that premise. Whether it is autonomous driving accidents, algorithmic discrimination, or unauthorized actions by AI agents, existing rules make it difficult to assign responsibility precisely. Global legislators face an urgent task: how to establish a new set of rules that can at least clarify responsible parties while preserving the momentum for innovation?

VI. The Climate Paradox: AI's Green Mask

More alarming is a recent study finding that AI's potential climate benefits are being offset by its role in boosting fossil fuels. AI could optimize energy dispatch and accelerate the discovery of new materials, but its own staggering electricity consumption is pushing power grids to lean further toward fossil energy. Not to mention that AI is also being used by oil and gas companies to improve exploration efficiency, indirectly prolonging the lifespan of fossil fuel extraction.

This paradox reveals the inherent limitations of technological solutions: AI is not a neutral tool; its direction and use depend on humanity's political and economic choices. When the narrative of "green AI" masks the energy geopolitics behind it, and when AI's expansion comes at the expense of climate goals, the sustainability of such "development" itself deserves repeated scrutiny.

VII. Music, Culture, and Social Erosion Elsewhere

Spotify's decision to distinguish between AI-generated artists and human artists and to stop recommending AI works, on the surface, caters to music producers' protests, but in reality reflects a deeper cultural anxiety. "I feel like I'm at war"—a musician described the confrontation with machine-generated music this way. AI is reshaping the underlying rules of cultural production: when music can be generated infinitely, how much value remains in originality? When algorithms can replicate style, where is the place of the "author"?

These seemingly marginal cultural frictions are actually a microcosm of society's overall adaptation to the impact of AI. As AI becomes deeply embedded in human creative territory, we not only need to redefine "artist," but also to rethink the meaning of "unique human contribution."

Conclusion: From Hype Cycle to Governance Cycle

Over the past decade, we have completed a full AI hype cycle—from the "deep learning revolution" to the "generative AI frenzy," from "changing the world" to "reconstructing everything." Now, the hype is receding, and the gravity of reality is beginning to show.

AI is no longer just a future myth written by tech journalists; it has become a hub where political, economic, legal, moral, cultural, and climate issues intersect. We are entering a "post-hype era," marked neither by technological stagnation nor by utopian leaps, but by sharp friction between technological maturity and institutional lag.

For global decision-makers, the urgent priority is no longer asking "what can AI do," but confronting "who should be responsible for AI," "how can society coexist with AI," and "which boundaries must not be crossed." These questions have no easy answers, but the cost of ignoring them will inevitably be settled more fiercely in the future.From Edinburgh to Boston, from second-hand bookstores to data center construction sites, humanity is learning, in a clumsy but ultimately unavoidable way, to negotiate with this partner more powerful than any tool. And the outcome of this negotiation will ultimately determine the basic contours of the global order in the 21st century.