The Historical Rebuttal of the 2028 "AI Doomsday Theory"

The 2028 AI doomsday prophecy is a perfect brainstorming session, but the real economy is a super chaotic system. History repeatedly proves that when a vision derived from logic is too extreme, predictors often underestimate humanity’s seemingly inefficient but actually highly resilient adaptive capacity.

At every historical juncture, many experts have expressed bold expectations for the future: Keynes’ 1930 “Economic Possibilities for our Grandchildren,” Russell’s 1932 “In Praise of Idleness,” Soros’ 1987 “Productivity Paradox,” and Greaber’s 2013 “On the Phenomenon of Bullshit Jobs” — top social scientists in the halls of human wisdom have all demonstrated one point: In the face of major technological revolutions, humans often know little about the future social development path.

2026 is a key year for understanding how AI will impact macroeconomics. In the “looking forward” process, more subjective judgment needs to be nested within the logical foundation. This is also a major reason why “2028 AI Wasteland Literature” is so attractive — bold enough, but somewhat “black-and-white.”

However, we are surprised to see that the pessimism triggered by this “semi-open” thought experiment is spreading widely. This report mainly combines the “historical outlook” of the four sociologists mentioned above, from a broad social science perspective, to reconsider the 2028 AI doomsday prophecy.

Admittedly, we may be experiencing a moment of the 21st century “Luddite Fallacy,” but from the vantage point of these giants’ shoulders, we believe there is no reason to be overly pessimistic about the AI era — we should not underestimate humanity’s seemingly inefficient but highly resilient adaptive capacity.

1. John Keynes — “Economic Possibilities for our Grandchildren”: Infinite Human Desires

Essentially, Keynes’ 1930 “Economic Possibilities for our Grandchildren” and the AI doomsday prophecy both advocate a form of “technological determinism.”

In 1930, Keynes concluded — based on compound interest and technological progress — that by 2030, living standards would increase 4-8 times, and humans would only need to work 15 hours per week. This is very similar to Citrini’s AI doomsday theory, which states that AI will take over most jobs, leading to mass unemployment (and potential economic crises).

But this technological determinism greatly underestimates the expansion of human desires. Once basic needs are met, humans will create newer, more expensive, and more “wasteful” needs. Although AI eliminates “old jobs,” human desires will instantly fill the void, creating thousands of “new jobs” we cannot even name now — the destruction may be significant, but the rate and scale of creation are endless.

The AI doomsday view echoes Keynes’ most famous statement in the text: “When the economic problem has been solved, mankind will face its most real and permanent problem: how to use its leisure.” (Thus for the first time since his creation man will be faced with his real, his permanent problem—how to use his freedom from pressing economic cares, how to occupy the leisure…)

However, it’s important to recognize that a century has passed, and most people have not widely recognized the significance of leisure like Bertrand Russell did, nor do they have more leisure. Instead, through “David Greaber-style bullshit jobs” and “Keynesian consumption upgrades,” they have successfully kept themselves busy. Some workers even work “every day,” not just “15 hours a week” — all built on the backdrop of significantly increased productivity.

In other words, we have not developed more “objective capacity” for creation; rather, we are investing more toward the next milestone. Based on this, the 2028 AI doomsday prophecy essentially assumes humans will suddenly stop fussing and passively enjoy leisure — which does not align with social development logic.

Referring to Nobel laureate Amartya Sen’s 1998 work “Development as Freedom”: “Freedom” and “equity” are not luxuries of economic development but its core drivers and ultimate goals — if AI, by destroying first, deprives humans of “economic conditions” and “social opportunities,” such development will be difficult to sustain smoothly.

2. Bertrand Russell — “In Praise of Idleness”: Leisure as a “Civilization Asset” Not a “Political Debt”

“In Praise of Idleness” is more like the ancestor of the 2028 AI doomsday prophecy, but with a relatively mild tone. Simply put, Citrini describes what would happen if we ignore Russell’s advice in the “AI era”; but in reality, over the past decades, the world has never progressed as “In Praise of Idleness” predicted, and at the same time, human living standards have still experienced qualitative leaps.

Russell and Citrini Research share the foundation that technological progress has significantly reduced the necessary labor time to sustain survival.

“In Praise of Idleness” presents a thought experiment: suppose a factory invents new technology, doubling efficiency. Russell believed that wages should be halved, and everyone would then work only 4 hours a day. But in reality, over the past decades, the common response has been to lay off half the workers or keep them working 8 hours, leading to overcapacity, financial volatility, and unemployment.

The core behind real-world decisions is that humans are constrained by an outdated moral shackles of “work is virtuous,” leading half the population to overwork and the other half to unemployment and hunger. But the AI doomsday theory is overly worried that, under current distribution systems, the unemployed do not gain “leisure,” but rather “purchasing power” — leisure as “political debt,” with passive unemployment creating social bad debts that require income from AI taxes to address.

However, leisure is a “civilization asset,” not a “political debt.” Technological progress can free humans from trivial physical and mental labor, turning leisure into creativity, science, and art.

Human society is an “entropy-increasing” system, and society is not a machine that can be managed programmatically — in facing the “AI doomsday” brainstorm with 10% unemployment, political systems will not sit idly by but will respond through fiscal expansion or shortening legal working hours (such as a 4-day workweek). This spontaneous adjustment is overlooked by the doomsday theory.

Furthermore, equating declining labor participation with institutional chaos and demand contraction is a typical “labor equals survival” mindset. The productivity leap caused by AI itself is neutral; if existing production relations collapse, it’s not entirely AI’s fault but a disconnect between “moral remnants of labor” (more work gets more rewards, no work gets none) and modern productivity.

The fear of the AI doomsday is essentially the inability to imagine a distribution system that does not equate human worth with employment (providing labor) — anchoring human meaning and economic stability solely on employment.

Of course, both “In Praise of Idleness” and the AI doomsday theory have their issues: Russell underestimates the necessity of competition for human evolution, and Citrini overestimates the instantaneous penetration of technology into social structures — both suggest that social resilience is stronger than imagined.

3. Robert Solow — “The Solow Paradox”: Production Relations Have Inertia

“The Solow Paradox” is the target of fierce criticism from the AI doomsday theory. It is because of “The Solow Paradox” that we think technological progress is gentle when “learning from history.” The doom theory essentially warns that this “invisibility” is continuously accumulating energy, and once released, it will cause huge shocks.

But our current situation is that, before witnessing a productivity singularity, some are already pricing the end of certain business models — this is no longer a question of moats but whether the water source of this river still exists.

The core logic of the AI doomsday theory is “AI replacement speed is extremely fast, while society’s adaptation speed is very slow, leading to cliff-like unemployment and deflation effects.” But the Solow Paradox and the past three years’ reality tell us that the process of transforming technology into productivity remains relatively slow, and the so-called “doomsday” will be partly offset by “lag effects.”

The impact of AI on employment and production relations is less significant than macro factors and pandemics. In reality, enterprises are not just production units but complex interest-bargaining entities. Perhaps we are experiencing a 21st-century “Luddite Fallacy” moment — just as textile workers historically destroyed machines to resist automation, it’s reasonable to speculate that rapid AI development might also face systemic resistance in some industries.

We have long emphasized the decline of historical comparability. Besides traditional investment growth rates and proportions, there is also a shift in AI’s role: from “assistive tool” to “independent production unit.” But the responsibility and authority in production have significant inertia. Especially for American companies, with historically high profit margins, even “being cautious” may lack the urgency for transformation.

Extremely speaking, even if production relations change, “human participation” itself may become a scarce asset and a source of premium. Industrialization produced cheap textiles but also spawned luxury goods and “artisan spirit” with high premiums. Society will enter a new “reputation and service” economy, where human labor shifts from “physical/computational” to “emotional/trust-based” domains — sincere humans will become more “important,” not “insignificant.”

4. David Greaber — “Bullshit Jobs”: Ending Meaningless Jobs

“Bullshit Jobs” directly counters the AI doomsday theory: if social systems are inherently creating many “meaningless” jobs to maintain stability, then AI might instead trigger efficiency back to its roots, not cause societal collapse.

The AI doomsday theory is built on a fragile assumption: that all jobs have social value and are indispensable, so losing a job equals losing a value anchor and distribution rights.

But Greaber argues that modern capitalism and existing technological progress are actually used to push society to work harder (not less), creating many pointless jobs. These jobs produce no real social output; their role is to distribute purchasing power and consume time. Even if all these jobs disappeared tomorrow, the world wouldn’t be worse — it might even be better. If AI replaces these roles, it does not destroy human “productivity,” it just bursts the “administrative inflation” bubble.

This “doomsday” is actually a return to efficiency, not a decline of civilization. Of course, between efficiency and fairness, a strong government is needed to maintain basic fairness.

On the other hand, the advent of AI actually provides an opportunity for “reallocation due to technological reasons,” where the starting line provided by AI technology is “relatively” fair compared to previous endowments. AI’s “technological unemployment” forces society to shift from “job-centered” to “people-centered” approaches; an extreme version of this is the universal basic income (UBI).

The reason we think AI will bring “doomsday” is because we lack a new value anchor — AI does not create crises; it simply ends many “bullshit jobs” and shatters the collective belief that “everyone must be busy” (or that “work is virtue”).

In short, Citrini Research’s 2028 AI doomsday prophecy is a perfect brainstorming, but the real economy is a super chaotic system.

History repeatedly proves that when visions derived from logic are too extreme, predictors often underestimate humanity’s seemingly inefficient but highly resilient adaptive capacity. Only when leisure can be massively released will AI technology be able to match the social contributions of the three previous “industrial revolutions.”

Risk warning: The development speed of AI-related capabilities exceeds expectations; overseas economies face increasing populist political tendencies, causing global economic growth to slow unexpectedly; breakthroughs in AI technology lead to significant reductions in manufacturing costs, surging credit demand, and a new wave of productivity liberation.

Source: Guojin Securities

Risk warning and disclaimer

Market risks exist; investments should be cautious. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should consider whether any opinions, viewpoints, or conclusions herein are suitable for their particular circumstances. Invest accordingly at their own risk.