An analysis of US payroll data across 730+ occupations: employment among workers ages 22 to 25 in highly AI-exposed jobs is now shrinking by 3.8% per year
Last August, a team led by Stanford economist Erik Brynjolfsson published a deep look at the impact of AI on jobs, boosted by a …
Context & Ripple Effects
Related coverage has been converging on the same labor-market fault line: Stanford researchers previously reported a 13% decline in entry-level employment in the most AI-exposed fields since 2022, while later BLS-based reporting found AI-exposed occupations underperformed the broader labor market.
The new payroll analysis narrows that pattern to workers ages 22 to 25 and gives it a current annual rate. It also sits alongside reporting that productivity accelerated as AI-exposed sectors cooled entry-level hiring, linking hiring changes to a broader shift in how firms capture AI-related gains.
First-order effects
- Workers ages 22 to 25 in highly AI-exposed jobs face a shrinking pool of employment, making the earliest rung of affected career ladders harder to access.
- Employers in those occupations can reduce junior hiring while maintaining output, consistent with the reported cooling of entry-level hiring in AI-exposed sectors.
Second-order effects
- A weaker entry-level pipeline can push new graduates and early-career workers toward less AI-exposed roles, increasing competition for jobs where AI is not yet reshaping staffing as directly.
- If productivity gains arrive alongside reduced labor income, the immediate distributional effect is a larger share of returns flowing to capital rather than to junior workers' wages and employment.
Third-order effects
- If the pattern persists across more occupations, firms may redesign professional training around smaller junior cohorts and greater reliance on AI-assisted work, altering how expertise is developed inside organizations.
- The divergence between labor and capital income could make the fiscal and policy consequences of AI adoption more salient, including for tax bases that depend heavily on labor income.
The trend: AI adoption is increasingly showing up not simply as job displacement, but as a restructuring of entry-level hiring and the distribution of productivity gains between labor and capital.