Christopher L. Riley / Wikimedia Commons / CC BY-SA 4.0 / cropped
CLEVELAND – Workers in occupations more exposed to artificial intelligence are taking longer to find jobs, particularly younger workers, according to research published Monday by the Federal Reserve Bank of Cleveland. The study found no consistent evidence that those workers were leaving existing jobs at a faster rate.
The findings point to a hiring problem rather than a clear wave of AI-driven layoffs. For workers ages 23 to 54, unemployment rose roughly one percentage point more in the most AI-exposed occupations than in the least exposed, relative to a 2015-2019 baseline.
The November 9 Economic Commentary, “Job Finding and Unemployment in the Age of AI,” compares unemployment in May 2025 through April 2026 with that earlier baseline. Its analysis of movements into and out of jobs uses January 2024 through April 2026, a longer window needed to follow enough workers across consecutive monthly surveys.
Two task-based measures showed that monthly job-finding rates declined by roughly six percentage points more in the most exposed occupations than in the least exposed. That is a difference in the probability that an unemployed person finds employment the next month, not a 6% loss of jobs.
Against an average monthly job-finding rate of around 34%, the paper says a six-percentage-point decline implies average unemployment duration rising from about 10.5 weeks to more than 13 weeks. A third measure, based on actual AI use rather than potential task exposure, showed a similar but weaker relationship.
The age split was sharper for unemployment. Among workers ages 23 to 30, fitted estimates implied a gap of about two percentage points between the least and most exposed occupations. The corresponding gap for workers ages 45 to 54 was at most one percentage point.
By contrast, the results for job separations were mixed and not statistically robust. One exposure measure showed essentially no relationship, while the other two pointed in opposite directions. The authors interpret the overall pattern as consistent with a “no hire, no fire” labor market in which employers reduce new hiring rather than broadly replace existing workers.
The research uses the Current Population Survey to follow civilians ages 23 to 54 across months and sorts occupations by how closely their tasks overlap with AI capabilities. Unemployed workers are classified by their most recent occupation. That means the analysis misses new entrants who have never held a job and cannot fully capture people searching in a different field.
The results are observational associations, not proof that AI caused the changes in hiring or unemployment. The authors describe their analysis as an early, incomplete view of a longer adjustment. Their views do not necessarily represent those of the Cleveland Fed or the Federal Reserve Board.
JBizNews Desk | Cleveland
© JBizNews.com. All rights reserved. Unauthorized reproduction or redistribution is strictly prohibited.
