The AI Layoff Trap: Too Much AI, Too Few Jobs Could Hurt The Economy, Warn Economists

As companies across the globe race to plug artificial intelligence into everything from coding to customer support, a new research paper is throwing in a reality check. It suggests that going all-in on AI layoffs might not just hurt employees, it could backfire on businesses themselves. A March 2026 paper titled “The AI Layoff Trap” by economists Brett Hemenway Falk from the University of Pennsylvania and Gerry Tsoukalas of Boston University argues that firms may be stepping into what they call an “automation arms race.” When too many workers are replaced by AI, fewer people are left with incomes to actually buy products and services. Over time, that shrinking demand can hit companies where it hurts the most, revenue. As the paper puts it, “if

AI displaces human workers faster than the economy can reabsorb them, it risks eroding the very consumer demand firms depend on.”Why Companies Still Push For AutomationSo if the risks are known, why aren’t companies slowing down? According to the researchers, it comes down to competition. Each firm benefits directly from cutting labour costs, but the downside, reduced consumer spending, is shared across the entire economy. In a crowded market, no one wants to be the only player holding back. Economists call this a “negative externality.” In simpler terms, companies enjoy the savings but don’t fully feel the damage they help create. The result is what the authors describe as a “demand externality trap,” where businesses keep automating even when it may not be good for the bigger picture.Signs This Is Already HappeningThe paper points to early real-world signals. For instance, financial tech firm Block reportedly cut nearly half of its 10,000 employees in February 2026, with CEO Jack Dorsey suggesting AI made many roles unnecessary. On a broader scale, over 100,000 tech layoffs were recorded in 2025, with AI playing a role in more than half those cases. Roles in customer support, operations, and middle management have been particularly vulnerable.Why Popular Fixes May Fall ShortInterestingly, the researchers aren’t convinced by commonly suggested solutions like universal basic income (UBI), reskilling programs, or even worker equity models. These, they argue, deal with the aftermath rather than the root cause. Even with safety nets in place, companies still have strong incentives to replace workers with AI. So the cycle continues.Infosys CEO Salil Parekh Says AI Could Make Mid-Level IT Roles The New Power CentreThe One Solution That Might WorkThe paper suggests a more direct fix: a Pigouvian automation tax. In simple terms, this would mean taxing companies for replacing workers when it harms the broader economy. Think of it like a carbon tax, but for jobs. If automation reduces overall demand, firms would need to account for that cost too. The conclusion is blunt: “Only a Pigouvian automation tax can.”A Bigger Question About The FutureThe researchers aren’t anti-AI. Far from it. But they do raise an important question, what happens if automation moves faster than people can adapt? History shows that technology creates new jobs over time. But if the transition happens too quickly, we could end up in a strange situation where productivity rises, but fewer people can afford to participate in the economy. In other words, companies might save money today by replacing workers, but tomorrow, they could find themselves with fewer customers to sell to.