What A.I. Exposes About Capitalism

I recently wrote a piece arguing that we need to change the incentive structure around artificial intelligence before the race itself makes dangerous choices seem unavoidable. One reader, Prasidha, left a comment that seemed to ask the deeper question beneath the piece (the one I should have addressed): If A.I. exposes a recurring failure in capitalism, what should come after the version of capitalism we have now?

Her point was not that A.I. is uniquely corrupting. The opposite, really. A.I. belongs to a much longer history of technologies that produced real material gains while leaving much of their damage for society to absorb. Industrial capitalism brought abundance, but also pollution, dangerous labor conditions, and the destruction of older ways of life.

The digital platform economy brought convenience, connectivity, and new forms of enterprise, but also surveillance, dependency, attention extraction, and the weakening of local commercial life. A.I. may bring astonishing productivity, scientific acceleration, and new tools for creativity, but it also threatens to destabilize work, education, status, and the relationship between human usefulness and economic reward.

Prasidha framed the matter through Francis Fukuyama’s argument about liberal democracy, capitalism, and thymos (the human need for recognition). Liberal democracy, on Fukuyama’s account, satisfies the desire to be recognized as equal among other human beings. Free-market capitalism reinforces that settlement because it raises living standards and gives people a way to improve their condition. There is something powerful in this argument. Capitalism does not merely produce goods. At its best, it produces the sense that one’s birth need not determine one’s fate, that talent and effort might alter the shape of a life.

This is one reason capitalism has been morally resilient even when it has been economically brutal. It has often been defended not only as a machine for wealth creation but as a system of possibility. A person may start with little and still rise. A family may become more secure than the generation before it. A child may move farther than her parents. Even when the promise has been unevenly distributed, it has remained politically potent because it speaks to something deeper than consumption. It speaks to dignity.

But this is also where capitalism becomes fragile. Once recognition is routed primarily through the market, a person’s social worth begins to depend on employability, productivity, credentials, salary, and proximity to growth. A.I. threatens this arrangement because it may weaken the connection between labor and recognition. If machines can perform more of the cognitive work through which people earn income, status, and purpose, then the old bargain begins to fray. A stipend, whatever its practical merits, will not be enough; even a materially comfortable society could become politically unstable if large numbers of people come to experience themselves as unnecessary.

Human beings do not only want to consume. They want respect, usefulness, membership, and a plausible path upward. They want to feel that their effort enters the world and changes it. They want to be needed by other people, not merely maintained by institutions. This is why universal basic income, or U.B.I., has become both attractive and politically revealing. Its basic premise is that every adult receives a regular cash payment from the state, regardless of employment status, poverty status, or moral deservingness. Its proponents range from Silicon Valley technologists worried about automation, to labor figures like Andy Stern, to anti-poverty writers like Annie Lowrey and Rutger Bregman, to political figures and entrepreneurs such as Andrew Yang and Chris Hughes. They see in U.B.I. a way to simplify the welfare state, cushion workers against technological displacement, reduce poverty, and give people enough security to leave bad jobs, care for family members, retrain, create, or simply breathe.

On a practical level, the case for universal basic income is serious. A guaranteed income can make people less desperate, reduce bureaucratic cruelty, and give them enough security to leave bad or abusive jobs.

But it should not be mistaken for a magic solution. A check can keep someone alive; it cannot, by itself, give that person a role. It can relieve poverty without repairing status. It can soften inequality without changing who owns the machines, who sets wages, who controls institutions, or who gets to imagine a future. The danger is that U.B.I. becomes the humane face of exclusion: a settlement in which a small class builds, owns, and governs the productive systems of society while everyone else receives enough money to remain quiet. That may solve a budgetary problem, but it does not solve the problem of dignity. A society that solves the income problem while ignoring the recognition problem will discover that it has not solved the political problem.

This is why A.I. cannot be treated merely as another productivity tool. It touches the psychological foundation of modern capitalism. For two centuries, the basic promise has been that technological progress may disrupt particular jobs but will ultimately create new ones, raise living standards, and expand opportunity. That story has been true often enough to remain credible, but A.I. places pressure on the story because it is aimed not only at muscle but at judgment, language, analysis, design, coding, tutoring, writing, planning, and persuasion. It reaches into the categories of work that many people were told would protect them from automation.

Perhaps the most extreme predictions will prove wrong. Perhaps A.I. will augment more workers than it replaces. Perhaps new professions will emerge, as they have before. But the political significance of this moment does not depend on the most apocalyptic scenario coming true. What matters is that many people can now imagine a future in which the economy grows while their own bargaining power declines. They can imagine a world in which productivity rises while their children’s prospects narrow. They can imagine, with growing plausibility, that the ladder is being pulled up just as the machinery becomes powerful enough to make the people at the top richer than anyone in history.

That perception alone is dangerous. Liberal capitalism depends not only on output, but on legitimacy. People will tolerate inequality when they believe the system remains open, when they believe effort can be rewarded, when they believe today’s hierarchy is not tomorrow’s caste. Once they begin to feel that the game has been engineered against them, politics changes. Conspiracy becomes attractive, and violence becomes easier to justify. The defense that “the economy is growing” will not persuade people who experience that growth as their own dispossession.

This is why the next order of capitalism has to be designed around a broader account of value. Profit is a useful signal, telling us what people want, where demand exists, and where resources may be flowing efficiently. But profit is an incomplete signal. It does not know whether a profession is being hollowed out or whether children are being educated well. It does not know whether communities are weakening, whether families are stable, whether local institutions are dying, whether citizens still trust one another, or whether people feel they have a future.

A capitalism organized only around profit will treat these things as externalities until they become emergencies. Then, once they become emergencies, the state is asked to repair what the market was never required to count.

A better capitalism would still reward invention, risk, and efficiency, but it would force powerful firms to internalize more of the social costs they create. In the A.I. case, that might mean taxes on automated labor or excess productivity gains, public equity stakes in frontier A.I. firms, data dividends, worker transition funds, mandatory impact assessments, stronger antitrust enforcement, and public investment in forms of work the market routinely undervalues (for example, teaching, nursing, child care, elder care, mental health, civic administration, local journalism, environmental repair).

The exact mechanism matters less than the principle. If a company becomes valuable because it has drawn on public research, public infrastructure, copyrighted human culture, user data, labor-displacing automation, and a legal system that protects its property, then the public should have a claim on some portion of the upside.

We already accept this logic in other contexts. Companies are not allowed to dump poison into rivers simply because doing so is profitable. They are not allowed to sell unsafe drugs simply because there is demand. They are not allowed to build factories without labor standards simply because lower wages improve margins. Far from ending capitalism, these rules made the system politically survivable. They reflected the recognition that markets require boundaries because human beings are not only consumers; they are workers, neighbors, parents, patients, citizens, and future ancestors.

A.I. should be governed in the same spirit. Before systems are deployed at enormous scale, companies should have to answer basic questions about labor effects, educational effects, privacy, safety, dependency, and concentration of power. The burden should not fall entirely on society after the fact. We should not wait until a profession has been destroyed, a school system has been distorted or a public sphere has been flooded with synthetic persuasion before asking whether the gains were worth the damage.

This does not mean freezing technology in place. The point is to reject the superstition that speed is neutral and that whatever can be built fastest should become the foundation of everyone else’s life. A society capable of governing technology would distinguish between invention and deployment, between possibility and permission, and between private advantage and public good.

The deeper challenge is recognition. If A.I. reduces the market value of some forms of human labor, then society will need to build other durable sources of dignity. But these cannot be merely rhetorical. We cannot praise teachers while underpaying them, celebrate care workers while exhausting them, admire parents while making family life economically impossible, or invoke community while allowing every local institution to be stripped for parts. If care work is valuable, it has to be funded. If education matters, educators have to be treated as central rather than ornamental. If civic life is necessary, the people who maintain it need more than ceremonial gratitude.

This is where Prasidha’s question becomes most important. Can we design a capitalism that continues to raise living standards while refusing to make profit the sole organizing principle of society? I think we can, but only if we abandon a certain naïveté. The old faith was that private incentives, left largely to themselves, would generate public benefit as a byproduct. Sometimes they do. Often they do not. The history of capitalism is not a simple story of exploitation, nor is it a simple story of liberation. It is a story of extraordinary creative force repeatedly outrunning the institutions meant to civilize it.

A.I. gives us the chance to see this pattern while there is still time to act. Unlike earlier technological revolutions, this one is unfolding under conditions of unusually high public awareness. People are not waiting a century to notice the factory smoke; they are already asking what happens to work, school, art, truth, privacy, politics, and human status. There is anxiety, but there is also political intelligence in that anxiety. It is a form of early warning.

The danger is that this awareness curdles into fatalism. Some will say the race cannot be slowed. Some will say China requires us to move faster. Some will say the market has already decided. Some will say the technology is inevitable, as if inevitability were a natural fact rather than a story powerful people tell about choices they prefer not to defend. The future will be shaped by A.I., but the terms of that shaping remain political. The question is who gets to decide them.

A post-naïve capitalism would begin from the recognition that markets are tools, not gods. It would preserve what capitalism does well: invention, experimentation, ambition, distributed problem-solving, the capacity to generate abundance. But it would stop pretending that abundance automatically produces legitimacy, or that efficiency automatically produces dignity. It would measure success not only by what is built, but by what is sustained: families, schools, professions, towns, trust, agency, public competence, and democratic control.

The purpose of such a system would not be to make capitalism benign. Capitalism is not benign. Its energy comes partly from competition, ambition, and dissatisfaction. The task is to discipline that energy toward human ends. A society should be able to say that some efficiencies are too destructive, some concentrations of power too dangerous, some forms of extraction too corrosive, and some kinds of abundance too detached from human flourishing.

A.I. is forcing this question because it makes the old evasions harder to maintain. If the technology succeeds on its own terms, it may generate wealth at a scale that exposes the poverty of our moral vocabulary. What does it mean for a society to become richer if its people feel less needed? What does progress mean if it dissolves the conditions under which ordinary people experience themselves as agents? What is the point of raising living standards if the structure of recognition collapses beneath them?

The answer cannot be nostalgia. We cannot return to a pre-A.I. economy, just as industrial societies could not return to the village. But neither should we accept a future in which every social priority is rearranged around the profit motives of a few firms. The real opportunity of this moment is to insist that technological power be matched by institutional imagination.

A.I. may become the technology that finally forces capitalism to grow up: to admit that profit is indispensable but insufficient, that innovation is powerful but morally incomplete, and that human beings need more than cheaper goods and faster systems. They need dignity and a place in the world.

The task, then, is to build a society in which powerful technologies are governed not only by the ambitions of those who create them, but by the claims of the people who must live with their consequences.