Artificial intelligence (AI)-led layoffs may not just be a labour market problem, they could end up hurting companies themselves, according to a new academic study that warns of a self-defeating “automation trap”.

The paper, “The AI Layoff Trap” by researchers Brett Hemenway Falk (University of Pennsylvania) and Gerry Tsoukalas (Boston University), argues that firms may be caught in a competitive cycle of excessive automation that ultimately erodes consumer demand. 

“At the limit, this becomes self-destructive: firms automate their way to boundless productivity and zero demand,” the paper said. 

Must read: Oracle lays off thousands globally, India cuts may reach 12,000

The study shows a basic problem with the AI boom. Companies use it to save money, but when people lose jobs, they have less money to spend. Since these same people are customers, overall demand in the economy starts to fall.

The problem, the researchers say, is not a lack of foresight. Even when companies understand the risks, competitive pressure forces them to continue automating.

Must read: No meeting, no warning, just a mail: Oracle layoffs 2026 raise questions on process, pay, and scale

“We show that knowing this is not enough for firms to stop it,” the study notes, explaining that companies are effectively trapped in an “automation arms race”. 

This dynamic creates what economists call a demand externality. Each firm benefits from lower costs when it automates, but only bears a fraction of the broader economic damage caused by reduced consumer spending. The rest of the impact is spread across competitors.

Must read: Meta layoffs 2026: The severance package offered to 8000 employees

The result is that firms automate more than is collectively optimal, even to their own detriment.

“The resulting loss harms both workers and firm owners,” the paper states, adding that the outcome is not just redistribution but a “deadweight loss” for the economy. 

Real-world trends appear to reflect this shift. The study points to over 100,000 tech layoffs in 2025 alone, with AI cited as a key driver in more than half the cases. So far in 2026, over 92,000 employees have been laid off across 98 companies. 

Must read: AI isn’t the real reason for tech layoffs? Salesforce CEO Marc Benioff calls it a ‘lazy way out’

Crucially, the distortion worsens as competition intensifies and AI becomes more powerful. “More competition and ‘better’ AI amplify the excess,” the study said. 

The research also evaluates popular policy responses to automation, from universal basic income (UBI) to reskilling programmes and worker equity participation, and finds that most fall short of addressing the core issue.

“Only a Pigouvian automation tax can,” the paper concludes, referring to a targeted tax on automation designed to offset its broader economic costs. 

Such a tax would effectively make companies internalise the demand loss caused by layoffs, aligning private incentives with the overall health of the economy.