How Retrainable are AI-Exposed Workers?

As artificial intelligence (AI) capabilities advance, will workers best adapt by reskilling into AI-complementary work or by sorting into occupations less exposed to AI? To answer this question, we assemble a large-scale dataset of occupational training spells funded by the U.S. Workforce Innovation and Opportunity Act from 2012–2024. We link pre- and post-training occupations to task-level AI exposure measures and estimate the returns to training by comparing trainees to matched workers who sought workforce services but received only job search assistance. While quarterly earnings returns are consistently high for workers from low AI-exposure occupations, returns for workers from high-exposure occupations rose sharply over our sample period—from about $1,000 per quarter before 2020 to $3,000 by 2022–2024. We attribute these gains primarily to transitions into less AI-exposed occupations and, to a lesser extent, to the expansion of training programs that build AI-complementary skills. To quantify when training into higher AI exposure work pays off, we construct a new AI Retrainability Index (AIR) and find that a large share of occupations are "AI-retrainable," pointing to broad potential for adaptation.