AI Layoffs Are Rising in Headlines, but Unemployment Is Holding Near 4.1%, Harvard Economists Say

The AI jobs apocalypse has been forecast loudly. It has not shown up in the national numbers.

The unemployment rate is hovering around 4.1%, largely buoyed by growth in healthcare, according to a Harvard Gazette interview published Oct. 20 with Harvard Business School's Joseph B. Fuller and Raffaella Sadun.

That sits next to a steady stream of layoff announcements with an AI label on them.

The scary numbers

News outlets have reported more layoffs attributed to AI this year than last. Anthropic CEO Dario Amodei said in 2025 that AI could wipe out half of entry-level white-collar jobs within five years.

Entry-level job postings have fallen by up to 73% over four years, by some counts. That is a big drop, though the figure depends on who is counting and how.

The reassuring numbers

For every article announcing AI-related layoffs, another dismisses the claim as "AI-washing." That is the idea that companies cut staff for ordinary business reasons and borrow the AI story because it sounds forward-looking.

Management consulting firm Accenture says 52% of top executives expect AI to increase entry-level hiring, not decrease it. That is an expectation from executives, not a hiring result.

Fuller and Sadun say companies still need human expertise for now. Their caveat is that roles may change soon as AI handles more tasks.

Why the data may be a poor guide

Doug Elmendorf, a Harvard Kennedy School economist and former director of the Congressional Budget Office, argues the answer takes patience.

"What we've seen so far in the labor market from artificial intelligence has very little predictive power for what we're going to see in the labor market because of AI in five years, or 10 years, or 15 years," Elmendorf said.

This spring Elmendorf, Karen Dynan of the Kennedy School and Louise Sheiner of the Brookings Institution wrote a National Bureau of Economic Research working paper. It lays out four scenarios for AI's effect on the economy.

They run from a moderate GDP boost with little reduction in the number of workers to much faster GDP growth with persistently high unemployment. The authors declined to say which is most likely.

Elmendorf was willing to give his own read in the interview. "I think it's very likely that there will be significant job losses from artificial intelligence in the coming couple of decades, but that most — not all, but most — of those people will be able to find new work," he said.

The China shock yardstick

The paper uses the early-2000s China shock as a comparison. Surging cheap Chinese exports devastated manufacturing communities across the Midwest and South.

Other researchers estimate that disruption cost roughly 1.5 to 2 million jobs between 2000 and 2007. The authors' AI scenario is far larger, with 3 million people, or roughly 2% of the labor force, out of work at any given time.

Those manufacturing towns got retraining promises and not much else. Whatever AI does to white-collar work, the China shock is the cautionary example of what slow, unfocused policy looks like.

What history can and cannot tell us

A New York Times analysis published Oct. 22 frames the same problem from another angle. Economists are hoping earlier episodes of labor disruption, from the Industrial Revolution to the Great Depression, can help the country prepare for AI.

The analogies are imperfect. Past disruptions hit factory floors and farms. AI is aimed at tasks done by college-educated workers, the group that has traditionally been most insulated.

The disagreement

Amodei's warning and the Harvard economists' caution are not the same claim. Amodei projected a steep loss of entry-level white-collar jobs within five years. Elmendorf says the current labor market tells us little about what comes in five, 10 or 15 years, and expects most displaced workers to find new work.

The layoff-skeptic camp points to the unemployment rate and to Accenture's survey. The alarmed camp points to the 73% drop in entry-level postings. Both sets of numbers are real.

The unresolved question is which of the NBER paper's four scenarios the economy is tracking. The authors have not ranked them, and the entry-level hiring data over the next few years is the clearest early test.

If entry-level postings keep falling while overall unemployment stays near 4.1%, the damage will be concentrated among young workers trying to get their first job. That is not something a national unemployment rate would show.