Stanford study shows AI-driven employment gap widening for young workers

Stanford University economists report a widening employment gap for 22-25 year olds in high-AI-exposure occupations, raising concerns about youth unemployment spikes and economic instability.

Stanford University economists published an August 2026 update to their study 'Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence.' The analysis reveals that for workers aged 22 to 25, employment levels in occupations with high AI exposure are 19 percentage points below those in less AI-exposed occupations. This gap has widened from 13 percentage points observed in the previous year's study. The research uses anonymized ADP payroll data and the Anthropic Economic Index to measure occupational AI exposure, categorizing it as either 'automative' (full replacement of human labor) or 'augmentative' (complementary to human labor). Occupations like accountants and auditors show high susceptibility to automative AI use, while chief executives and registered nurses demonstrate high augmentative AI usage. For 22-25 year olds, occupations with high automative AI exposure show declining entry-level employment rates.

The study's mechanism centers on how AI adoption patterns affect entry-level labor markets. It identifies that occupations requiring higher codified knowledge show slower entry-level employment growth for young workers, while those requiring tacit knowledge show faster growth for mid-career and senior workers. This distinction explains why the employment gap widens specifically for young workers in high-automative AI occupations.

This trend creates friction for youth economic stability. A 11 percentage point decline in entry-level employment for 22-25 year olds in the top 40% of AI-impacted occupations since 2022 threatens to spike youth unemployment, risking economic instability and social strain without targeted interventions. The friction is most acute for workers entering fields where AI could fully replace entry-level roles.

What to watch: How interventions might mitigate the employment gap in high-automative AI occupations for young workers. Caveats include the study's reliance on ADP payroll data as its primary employment metric and the Anthropic Economic Index's measurement through Claude model queries. The analysis focuses exclusively on entry-level employment trends for workers aged 22-25.

Source: Ars Technica