CFOs say AI layoffs will be 9x higher this year, but still affect few

The most powerful voices in artificial intelligence (AI) really want you to worry about your job — if you’re an office worker. Microsoft AI chief Mustafa Suleyman predicts that AI will make office jobs crumble within 18 months. Anthropic CEO Dario Amodei thinks entry-level roles in the field will be halved in a similar timeframe.

Even Federal Reserve Chairman Jerome Powell has warned that AI is quietly impacting the job market while job creation remains near zero. But behind closed doors, business leaders who control their companies’ headcount are actually telling a more nuanced story.

Also read: Can AI create mass unemployment? Genoa Capital bets on productivity gains

An ongoing study by the National Bureau of Economic Research found that in a survey of 750 CFOs (chief financial officers) of US companies, less than half (44%) say they plan any AI-related job cuts.

When the co-authors calculated what this represents for the economy as a whole, they found that just 0.4%, or about 502,000 positions out of a total of approximately 125 million, are expected to be eliminated this year. About half of these losses will come from office workers.

This 9-fold increase in relation to the 55,000 layoffs attributed to AI last year draws attention — but is still insignificant compared to the total workforce.

“It’s not the doomsday scenario for jobs that sometimes makes the headlines,” John Graham, co-author of the study and director of the Duke CFO survey, conducted in partnership with the Federal Reserve Banks of Atlanta and Richmond, told Fortune.

Furthermore, the study identifies a large difference between perceived and real productivity gains with AI, showing that expectations are greater than reality.

The researchers say this likely reflects a delay in actually generating revenue. This reported range is in line with what economists have been saying about AI’s productivity gains.

Goldman Sachs senior economist Ronnie Walker noted in early March that amid enthusiasm for AI investments, “we have not yet found a significant relationship between productivity and AI adoption at the level of the economy as a whole.”

It’s not just economists; Workers have reported that AI is actually making them less, not more productive, increasing pressure on their workflows, with time spent on some responsibilities growing by up to 346%

Delayed productivity gains and the Solow paradox

Economists have a name for this difference—and it dates back to the dawn of the personal computer era.

In a comparison with the delay in technological innovations observed at the beginning of the internet era, researchers turn to the Solow paradox, also known as the productivity paradox, to contextualize the current discrepancy between perceived and real productivity with AI.

Coined by 1987 Nobel laureate Robert Solow, the paradox reflects the realization that transformative technologies — such as computers or, in this case, AI — can be everywhere in everyday life and yet not appear in economic indicators.

“You can see the computer age everywhere but in the productivity statistics,” Solow said.

Graham said what executives are seeing today in terms of productivity is more wishful thinking than concrete fact. Companies see the potential of AI without matching financial results.

“Companies have invested and are realizing a lot of interesting things that they are starting to do or hope to do in the near future,” he said. “But that’s not showing up in revenue yet.”

The current state of AI-related layoffs

However, the study still represents a step toward greater job losses due to technology, lending some credibility to what technology executives have been saying publicly.

Employers reported about 55,000 layoffs attributed to AI in 2025, according to research firm Challenger, Gray & Christmas. This represents just 4.5% of all job losses last year.

Still, if the study’s numbers are correct, that would mean this year’s predictions call for a 9x increase in AI-related layoffs.

There have already been several significant AI-related layoffs reported by companies this year. Jack Dorsey’s Block cut about 40% of its workforce, or more than 4,000 employees, because of technology.

Australian-American financial services company Atlassian reduced 10%. And Meta is reportedly planning to cut 20% of jobs, while CEO Mark Zuckerberg reportedly creates an AI agent that is a clone of himself.

Furthermore, the job market is on hold. U.S. employers reported 92,000 job cuts last month, and the unemployment rate rose to 4.4%.

But the report also concludes that AI adoption could actually lead to an increase in hiring among smaller companies, further contradicting public statements from top industry leaders.

Many smaller companies, or those with fewer than 500 employees, have only recently begun to invest in AI, as they bear the majority of operational costs related to the technology, which slows adoption.

However, as adoption increases, these companies have indicated that they intend to expand hiring in technical roles. Additionally, larger companies plan to keep technical positions stable.

“If anything, small businesses are hiring a little in the technical area, which will offset (the losses) to some extent,” Graham said.

Still, the study only paints a short-term picture, making it difficult to entirely dismiss the more alarmist predictions made by some technology leaders about AI’s ability to replace office jobs.

“Who knows what will happen in 2028?” Graham said. “I’m not making the prediction that there will never be job losses two, three or five years from now because of AI.”