AI Researchers: AI Will Destroy Consumer Demand, Not Just Jobs
A new paper released today by researchers Brett Hemenway Falk of the University of Pennsylvania and Gerry Tsoukalas of Boston University delivers a stark warning to the corporate world: the rush to automate is not just a threat to workers, but a self-destructive strategy for the firms themselves. Titled “The AI Layoff Trap,” the study argues that the current wave of AI-driven layoffs is creating a demand externality that threatens to erode the very consumer base companies rely on for survival.
The core thesis is deceptively simple. When firms replace human workers with AI, they save on labor costs but simultaneously eliminate the wages that those workers would have spent as consumers. If automation outpaces the economy’s ability to reabsorb displaced workers, aggregate demand collapses. The result is a scenario where firms automate their way to boundless productivity only to find there is no one left with the purchasing power to buy their goods.
What makes the situation particularly alarming is that this outcome is not the result of irrational behavior or short-sightedness. According to the authors, the dynamic is a classic Prisoner’s Dilemma. Every executive can see the “demand cliff” ahead. Yet, in a competitive market, a firm that refrains from automating while its rivals proceed will face higher costs, lose market share, and eventually fail. Consequently, rational, forward-looking firms are trapped in an automation arms race, displacing workers well beyond what is collectively optimal.
“The fear that technology will displace workers is at least as old as the Industrial Revolution,” the paper notes, referencing historical precedents from Ricardo to Keynes. However, the authors argue that the current AI wave is distinct because the “reinstatement effect”—the creation of new tasks to offset displacement—is failing to keep pace. Early data supports this concern: over 100,000 tech workers were laid off in 2025 with AI cited as a primary driver, and companies like Block and Salesforce have already slashed thousands of jobs, citing AI efficiency.
The study rigorously tested various proposed policy solutions to break this cycle, finding most to be ineffective against the core structural incentive. Universal Basic Income (UBI), for instance, raises the floor for living standards but does not alter the marginal incentive for a firm to automate. Similarly, capital income taxes and worker equity participation fail to correct the specific margin where the externality resides. The authors also debunked the hope that “better” AI would solve the problem; instead, they identify a “Red Queen effect” where improved AI productivity actually widens the gap between private incentives and collective welfare, accelerating the race toward the cliff.
The only mechanism the mathematical model identifies as capable of stopping the spiral is a Pigouvian automation tax. This would be a per-task levy set equal to the uninternalized demand loss caused by each replacement. By forcing firms to pay for the demand they destroy, the tax would align private profit motives with the collective good, effectively implementing the cooperative optimum that competition prevents.
The implications are profound. The researchers conclude that the current trajectory is not a transfer of wealth from workers to owners, but a deadweight loss that harms both. Workers lose income, and firm owners lose customers. “Neither can capital income taxes, worker equity participation, universal basic income, upskilling, or Coasian bargaining,” the paper states. “Only a Pigouvian automation tax can.”
As the authors point out, the problem is not that firms are unaware of the risk. It is that the market structure makes restraint impossible without external intervention. With 80% of U.S. workers holding jobs susceptible to automation, the window for policy action may be narrowing. The paper suggests that addressing the aftermath of displacement is insufficient; policymakers must now address the competitive incentives driving the automation arms race itself.
For CEOs currently celebrating efficiency gains, the message is clear: in a race to the bottom of the demand curve, the winner is the last one standing in an economy with no buyers. The trap is set, and without a corrective tax, the only way out may be through a crash that leaves no one unscathed.
Link to Research Study
https://arxiv.org/pdf/2603.20617