External signal·Stanford Digital Economy Lab·Aug 26, 2025·Erik Brynjolfsson, Bharat Chandar, Ruyu Chen·15 min read
Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence
“a 13 percent relative decline in employment”
Summary
The foundational empirical study behind the entry-level-jobs alarm. Using high-frequency ADP payroll records, the authors document a roughly 13% relative decline in employment for early-career workers (ages 22-25) in the most AI-exposed occupations since generative AI's widespread adoption, while employment for older and less-exposed workers held stable or grew. Declines are concentrated where AI automates rather than augments tasks, persist after controlling for firm-level shocks, and show up primarily as employment rather than wage adjustments.
Predictions for the future of work
Presents early, large-scale evidence consistent with AI beginning to have a significant and disproportionate impact on entry-level workers — a leading-indicator ("canary") reading rather than a full-economy verdict. Implies the first rung of the career ladder is where AI's labour effect appears first.
Originally published by Stanford Digital Economy Lab · Aug 26, 2025
Read the original at Stanford Digital Economy Lab