A survey of 750 U.S. chief financial officers by Duke University and the Federal Reserve Banks of Atlanta and Richmond reveals that 44% plan AI-related job cuts this year. Roughly 502,000 roles out of 125 million — just 0.4% of the workforce — are expected to be eliminated, a ninefold jump from the 55,000 AI-attributed layoffs in 2025.
"It's not the doomsday job scenario that you might sometimes see in the headlines," said John Graham, director of the Duke CFO survey and co-author of the NBER working paper. Despite rising tech budgets — 75% of CFOs are increasing spend, with 48% planning double-digit hikes — executives say AI potential is "not really showing up yet in revenue," echoing the productivity paradox Nobel laureate Robert Solow identified during the early internet era.
The data also shows a split: smaller firms are hiring technical talent to accelerate AI adoption, while larger companies hold headcount flat. With unemployment at 4.4% and overall headcount growth expectations dropping from 6% to 2%, the labor market faces a slow squeeze rather than a sudden shock.
