AI layoffs may hurt companies too, not just workers: Study warns of ‘automation trap’

A recent academic study reveals that layoffs driven by artificial intelligence (AI) may not only pose challenges for the labor market but could also jeopardize the companies implementing them. The research paper, titled "The AI Layoff Trap," authored by Brett Hemenway Falk from the University of Pennsylvania and Gerry Tsoukalas from Boston University, warns of a self-defeating cycle of excessive automation that threatens to diminish consumer demand. The study highlights a critical flaw in the ongoing AI revolution. While companies adopt AI to reduce operational costs, the resulting job losses mean that consumers have less disposable income, which in turn decreases overall demand in the economy. The researchers emphasize that the issue isn't merely a lack of foresight; even when companies recognize the potential risks, they feel compelled to automate due to competitive pressures. The paper illustrates that understanding the negative consequences of automation is insufficient to halt its progression. Companies find themselves ensnared in what the authors describe as an "automation arms race." This phenomenon leads to a situation known as a demand externality, where individual firms benefit from reduced costs but only absorb a fraction of the economic fallout from decreased consumer spending. The broader impact, however, is distributed among competitors, resulting in collective harm. Consequently, firms may automate beyond what is beneficial for the economy, even harming their own interests. The study states that this loss affects both employees and company owners, leading to what is termed a "deadweight loss" for the economy. Evidence of this trend is evident in the tech sector, which saw over 100,000 layoffs in 2025, with AI cited as a contributing factor in more than half of those cases. As of 2026, over 92,000 employees across 98 companies have already faced layoffs, underscoring the urgency of the issue. The study also discusses various policy interventions aimed at addressing automation, including universal basic income (UBI), reskilling initiatives, and worker equity participation. However, the authors find that most of these solutions fall short of tackling the underlying problem. They propose a Pigouvian automation tax as a more effective remedy, which would compel companies to account for the economic loss tied to layoffs, thereby aligning their private incentives with the broader health of the economy. In summary, as competition escalates and AI technology advances, the potential for companies to exacerbate this cycle of automation increases, necessitating thoughtful policy responses to avert long-term economic detriment.