A new peer-reviewed economic study argues that unchecked AI-driven automation could, under specific conditions, create a self-reinforcing cycle of job losses and declining demand.

The paper, titled “The AI Layoff Trap” (March 2, 2026), is authored by researchers affiliated with the Wharton School at the University of Pennsylvania and Boston University.

It models a scenario where firms adopt AI to reduce costs, prompting competitors to follow, leading to widespread labor displacement.

Because workers also function as consumers, falling employment reduces overall demand, which in turn accelerates further automation and cost-cutting.

The authors describe this as a feedback loop in which individually rational corporate decisions collectively produce macroeconomic contraction.

The study evaluates multiple policy responses, including universal basic income, taxation, retraining programs, and coordination mechanisms, finding limited effectiveness in most cases under model assumptions.

It identifies a Pigouvian automation tax—levied per task replaced by AI—as the only mechanism capable of breaking the cycle within the model.

The paper concludes that without structural policy intervention, advanced automation could lead to rising productivity alongside weakening consumer demand pressures.