AI-related layoffs keep coming. But there’s more to the story

In 2025, employers cited artificial intelligence as the rationale for nearly 55,000 layoffs at companies like Amazon and Microsoft. And with the new year barely underway, we’re already seeing a new crop of AI-related job cuts. 

Citigroup is cutting over a thousand jobs, according to Bloomberg, and in a memo this week, CEO Jane Fraser warned of more layoffs later this year. “Over time, we can expect automation, AI and further process simplification to reshape how work gets done,” she added. Meanwhile, Meta is conducting more layoffs in its virtual reality division, cutting about 1,500 jobs as part of a broader strategic shift to invest further in AI. 

Given these reports, many observers have been quick to believe that workers are losing jobs to generative AI. But there is little evidence that automation is displacing workers en masse just yet—or even drastically changing how businesses operate. 

According to a recent analysis by the Brookings Institution and the Budget Lab at Yale University, the proportion of workers in jobs that are ripe for AI disruption has remained steady since ChatGPT launched in 2022. What’s more, there are all kinds of forces shaping the labor market right now, including changes in immigration policy that have curbed employment growth. 

“What we’re seeing overall right now is consistent with a labor market that has been hit with a lot of uncertainty in the macroeconomic environment,” says Martha Gimbel, executive director of the Budget Lab. “The immigration changes are making it really hard to interpret changes in the jobs numbers. And if you look for any signs of changes that seem to be due to AI, those are not yet showing up.” 

Still, a number of experts have pointed to AI adoption to explain the recent spike in labor productivity, which measures hourly worker output. In the third quarter of 2025, labor productivity climbed by 4.9%, the highest increase in two years. Some economists have speculated this is a sign that the growing adoption of AI across companies may in fact be boosting efficiency, despite the slow rate of hiring in 2025. 

But Gimbel argues productivity is too “noisy” a metric to accurately capture the impact of AI, particularly over just one quarter. “Productivity growth will be really high or really low in one quarter,” she says. “And if it fits their preferred narrative, people will jump on that.” A single quarter of high productivity should not be seen as a clear indicator of anything, she says, in part because labor productivity is imprecise and vulnerable to measurement error. That has been especially true in recent years because the pandemic threw a wrench in the system that is still being sorted out.  

“You had all these issues with productivity measurement in the pandemic because people largely fired low wage workers who tend to be less productive,” Gimbel says. “So you saw this huge jump in productivity, and then it came back down as those people were hired back. Was there actually a change in productivity in the economy? No.”

Research also shows that while AI might improve efficiency to some extent, it creates additional work that can hamper productivity. A new Workday report found that nearly 40% of the time saved by using AI is lost to rework; on average, workers spend 1.5 weeks annually correcting or otherwise fixing AI-generated content. 

As for whether AI is eliminating jobs, that’s not evident in jobs data just yet—and unemployment figures do not reflect any notable changes either. 

While the most recent jobs report does indicate a marked decline in employment across specific sectors, namely professional and business services, Gimbel says it’s too soon to say whether any of that is actually due to AI. She says it might take an economic downturn to really see that shift. “The place to start looking for the impacts of AI is when we have a recession,” she says. “That is usually when technological change really takes off.” 

All that said, Gimbel is closely watching sectors that have high adoption of AI, which includes not just tech, but also the arts and education. Even if concerns about AI usage in the workplace are overblown at the moment, workers will certainly start to feel the effects of it in the years to come. 

“It would be unusual for a new technology to have no impact on the labor market,” Gimbel says. “We just still need to find out how fast, and where.”