AI won’t only change our everyday lives, but it could also change how broader human society is organized.
In a stark warning about artificial intelligence’s societal implications, MIT economist David Autor has articulated a troubling vision of how AI could fundamentally undermine democratic governance. Speaking with characteristic directness, Autor frames the challenge not as one of material scarcity but as a crisis of distribution, identity, and political legitimacy. His concerns center on a deceptively simple mechanism: the wholesale devaluation of human labor in an AI-dominated economy.
“The bad scenario is one where the value of human expertise is just liquidated very rapidly, and that would create enormous problems for society, and it wouldn’t create problems of lack of material wealth,” Autor explains. “We’d have tons and tons of productivity and resources, but it would create a huge problem for distribution, for identity, and for our democracy.”
The crux of Autor’s argument rests on labor’s foundational role in contemporary society. “Most of our income distribution depends on labor, the value of labor. That’s how most of us get our income. So if all of a sudden labor were devalued by machinery, most of us would have nothing to sell to the market and therefore no way to capture all the resources that are being created.”
But Autor’s concerns extend beyond economics into the philosophical underpinnings of democratic participation. “Democracy depends on the idea that most people in the society are both claimants on the society and contributors to the society. And how do we contribute? We contribute with our work and we contribute with our taxation. And so if all of a sudden we had nothing to contribute from a market sense—I don’t mean from a moral sense—because our labor wasn’t any more valuable, but we were just claimants and we said, well, give me my UBI, I think that would be extremely a fundamental existential threat to the way our democracy works.”
To illustrate his point, Autor invokes an unlikely reference: Pixar’s WALL-E. “I don’t know how many of your viewers have seen the movie WALL-E. WALL-E’s a future where robots are doing all the work and people just float around in hovering armchairs. They weigh 300 pounds. They watch holographic TV and drink big gulps. They’re just sloths sitting in lounge chairs and it looks horrible. But whenever I watch that movie, I think, well, that’s the good scenario. That’s the scenario where they’ve actually figured out the income distribution problem, and then they just sit around doing nothing. It doesn’t look fun, but at least they’re not at war with one another.”
The alternative, in Autor’s view, is far grimmer. “I don’t think that’s the likely scenario. I think in a world in which all the resources were generated by machinery and only a few people own that—the owners of the AI companies, whatever—we would have a very unequal and very fractious society, and that’s how most of human history has been. So this is not just a dystopian fantasy. I think we want to preserve the labor market because we want to preserve democracy. We don’t have really great alternative models of this at present.”
Autor’s warning arrives at a moment of intense debate over AI’s labor market effects. Recent developments have vindicated some of his concerns: major corporations are increasingly deploying AI tools to augment or replace knowledge workers, from customer service representatives to software developers and content creators. The rapid adoption of generative AI systems like ChatGPT and Claude has accelerated anxieties about white-collar job displacement, extending automation threats beyond manufacturing into domains once considered immune.
His analysis also resonates with broader discussions about wealth concentration in the technology sector. The companies developing frontier AI models represent some of the most valuable enterprises in human history, yet their employee bases remain remarkably small relative to their market capitalizations. This capital-intensive, labor-light model of value creation exemplifies precisely the dynamic Autor warns against: immense wealth generation flowing to a narrow ownership class while the labor market’s capacity to distribute prosperity atrophies.
Perhaps most provocatively, Autor challenges the techno-optimist assumption that productivity gains automatically translate into shared prosperity. His insistence that democracy requires citizens to be contributors, not merely claimants, cuts against proposals for universal basic income as a solution to technological unemployment. In Autor’s framework, UBI might solve the distribution problem in narrow economic terms while hollowing out the civic foundations that sustain democratic legitimacy. The challenge, then, is not simply ensuring everyone has enough to consume, but preserving meaningful economic participation as the basis for political citizenship.