Anxieties about AI putting people out of jobs is bubbling over: A Reuters/Ipsos poll from August 2025 found 71% of Americans feared permanent job loss as a result of AI. Last week, Amazon announced 16,000 roles across the company would be slashed, adding to a total of more than 30,000 job cuts since October 2025. The move coincided with Amazon’s push toward AI development, though the tech giant attributed the reductions to an attempt to slash bureaucracy, not the technology.

A recent report from the Yale Budget Lab suggests there’s something to Amazon’s assertions that these mass cuts, even at tech companies, are not the result of AI displacing workers.

“While anxiety over the effects of AI on today’s labor market is widespread, our data suggests it remains largely speculative,” the report said. “The picture of AI’s impact on the labor market that emerges from our data is one that largely reflects stability, not major disruption at an economy-wide level.”

In order to measure AI’s impact on the labor force, the Yale Budget Lab tracked occupational mix, or changes in the types of jobs people in the U.S. have held, as well as length of unemployment for jobs with high exposure to AI.

While there have been changes in occupational mix since the 2022 release of ChatGPT, the rate of change has not increased enough to signal a massive shift, the report said. In addition, the length of unemployment for individuals with jobs that have high exposure to AI remained the same over time. Both metrics signaled no evidence of a massive labor disruption, from AI or another factor.

“No matter which way you look at the data, at this exact moment, it just doesn’t seem like there’s major macroeconomic effects here,” Martha Gimbel, executive director and cofounder of the Yale Budget Lab, told Fortune.

‘AI washing’ in action

The Yale Budget Lab’s assertion comes in the face of other pieces of data some have interpreted as a harbinger of massive labor disruptions. An MIT report released in November 2025 found current AI systems can already complete the tasks of nearly 12% of the workforce. Goldman Sachs predicted 6% to 7% of the U.S. workforce could be displaced if AI technologies become widely adopted.

Those forecasts don’t reflect today’s situation, despite growing concern over AI-related job losses. The disparity between AI anxiety around job displacement and the data indicating otherwise has led to concerns of “AI washing,” or the false attribution of AI to companies downsizing their workforces.

An Oxford Economics report last month backed up this idea, citing data from outplacement firm Challenger, Gray & Christmas: While 55,000 U.S. job cuts in the first 11 months of 2025 were attributed to AI, they represented only 4.5% of total reported job cuts. By contrast, job losses as a result of standard “market and economic conditions” totaled 245,000.

“We suspect some firms are trying to dress up layoffs as a good news story rather than bad news, such as past over-hiring,” the Oxford report said.

According to Yale Budget Lab’s Gimbel, one reason companies are attributing layoffs to AI is a way to avoid telling investors the company has had trouble navigating dwindling immigration, tariffs, and other policy uncertainties that inevitably shake up the labor force. AI-related anxiety has allowed the technology to become a convenient scapegoat for CEOs when it comes time to face skeptical investors.

“If you’re a CEO, what are you going to say? ‘Hi, I’m a really bad CEO. I totally mismanaged the macroeconomic situation for the last couple of years, so now a bunch of you are going to have to lose your jobs, but shareholders should keep trusting me moving forward?’” Gimbel said. “No, you’re not going to say that. You’re going to say, ‘The world is changing quickly, and we’re going to rightsize the company and make investments moving forward so we can be the most productive version of ourselves to win the future.’”

What’s really happening with the labor market?

She noted it’s much more realistic to attribute the low-hire, low-fire labor market conditions to the myriad political factors shaking up the economy as well as the aftereffects of the pandemic-era hiring surge and the Federal Reserve’s hiking cycle that naturally slowed down the job market.

To be sure, economic constraints could have an impact on how quickly new technologies are implemented, providing a blueprint for when AI could begin to more heavily impact labor, Gimbel suggested. During the first Industrial Revolution, for example, trade restrictions from the Napoleonic Wars led mill owners to rush to invest in technologies like the power loom and spinning jenny that automated weaving and displaced workers.

“Technological change does not happen in a vacuum,” she said.

AI’s next big test in the labor market will be if a recession comes, Gimbel said, necessitating changes that would incentivize mass adoption of AI. According to PwC data, AI adoption and productivity gains have been modest, with 56% of companies reporting they are getting “nothing out of” AI yet.

If there are big changes to the job market as a result of AI, Gimbel said it will be reflected in massive changes to the mix of jobs people hold and the length of unemployment for people with high exposure to AI in their previous jobs. Otherwise, she noted, it’s not time to sound the alarm.

“If you think the AI apocalypse for the labor market is coming, it’s not helpful to declare that it’s here before it’s here,” she said. “Any of this can change. That’s why we’re tracking it … Just because a technology can do something doesn’t mean that everyone loses their jobs tomorrow. It doesn’t mean they won’t lose their jobs in five years, though.”