According to a Brookings Institution study prepared for the National Bureau of Economic Research, 37.1 million U.S. workers are employed in jobs with the highest AI exposure. Of those, 26.5 million have above-median adaptive capacity, meaning they hold credentials, skills, or access to training that could support role transition. The remaining roughly six million do not. According to the study, women comprise 86% of the most vulnerable group.
The study draws a line between exposure and vulnerability that much AI displacement coverage flattens. Exposure is not destiny. A worker in a highly exposed role with strong adaptive capacity, the right credentials, employer-provided retraining, or transferable skills, is not the same as a worker in the same role without those resources. The study’s usefulness is in identifying where the policy intervention gap actually sits: not broadly among all AI-exposed workers, but specifically among the roughly six million whose exposure is high and whose adaptive resources are limited.
That distinction matters this week. Block has publicly attributed recent workforce reductions to AI automation, an explicit company statement that belongs in any honest accounting of how AI displacement is unfolding in practice. Specific headcount figures from that and other reported tech layoff announcements are pending source verification and will be updated when confirmed.
The Brookings/NBER framework gives enterprise HR teams, policymakers, and workforce development programs a sharper diagnostic than aggregate displacement estimates. The question isn’t how many workers AI will affect. It’s which workers face both high exposure and low adaptive capacity simultaneously, and what investments close that gap before the restructuring reaches them.