Behavioral economist Alex Imas warns AI could trigger demand collapse and negative growth

University of Chicago behavioral economist Alex Imas is challenging the conventional wisdom among economists that AI will ultimately create as many jobs as it displaces, arguing instead that it could lead to unprecedented economic disruptions like demand collapse and even negative growth. In recent discussions, including a Bloomberg Odd Lots podcast and an accompanying article, Imas questions the optimistic view that productivity gains from AI will naturally generate new demand and opportunities, as happened with past technologies like the steam engine. According to the podcast, while many economists expect short-term pain followed by balance, Imas warns this time might be different, with AI potentially causing spirals of reduced output and unemployment.

Imas outlines scenarios where AI automation triggers a demand collapse, drawing on New Keynesian economics. As reported in detailed analyses of his work, when large segments of the workforce lose income due to automation, consumer spending plummets because those affected have little money to buy goods and services. Firms, anticipating this drop, cut production, exacerbating the cycle—especially under nominal rigidities like sticky wages or prices that prevent quick adjustments. This differs from classic recessions, Imas explains, because AI could automate broadly and rapidly, hitting heterogeneous groups with varying marginal propensities to consume, potentially pushing economies toward a zero lower bound on interest rates where traditional stimulus fails.

Beyond immediate job loss, Imas highlights risks of diverging GDP and welfare in an AI-driven world. Simple narratives of "automation means abundance" overlook satiation—where humans reach points of satisfaction with goods—and behavioral factors like meaning and preferences that AI might not fulfill. In one model discussed in his interviews, widespread dissaving could trap societies in a high-tech, low-capital state, evoking dystopian visions like Isaac Asimov's Caves of Steel, where advanced technology coexists with economic stagnation. Even if GDP grows, welfare could suffer if productivity benefits accrue unevenly to capital owners rather than workers who can spend on scarce human-centric services.

Imas's research also reframes AI's impact on jobs beyond crude exposure metrics, such as studies claiming 80% of U.S. workers face some task automation. He argues that risk depends not on average exposure but on job structure: positions built around a few core, automatable tasks are most vulnerable, especially if firm incentives favor investment in low-dimensional automation. His studies on AI in creative work and principal-agent dynamics reveal how human traits influence machine interactions, potentially amplifying inequality through "machine fluency" gaps—where some people better direct AI, inheriting and transforming economic heterogeneity.

This perspective matters profoundly for policymakers, workers, and businesses as AI adoption accelerates. Millions in exposed sectors—from routine tasks to creative fields—could face displacement without the rebalancing seen historically, affecting global growth and inequality. What happens next remains uncertain: Imas expresses measured skepticism about productivity gains trickling down to labor, urging closer scrutiny of behavioral economics in AI scenarios. As debates intensify, his ideas, featured prominently on platforms like Bloomberg and his Substack, call for rethinking safeguards like retraining or demand supports to avert spirals.