−15.7%
▼ 15.7 pts since Nov 2022
estimated employment gap for workers aged 22–25, most vs least AI-exposed occupations, December 2025
20212025
What the chart shows
Brynjolfsson and colleagues group occupations into fifths by exposure to AI and compare young workers’ employment in the most and least exposed groups within the same firms. By December 2025 the estimate for workers aged 22–25 stood at −15.7% relative to November 2022, the lowest point in the series.
Until mid-2024 the estimate stayed between −6.2% and +1.3%. It was −0.1% in July 2024, fell in most months after, and passed −10% in March 2025. The AI Index describes this as a fall of roughly 16%.
Key numbers
| Measure | Value |
|---|---|
| Dec 2025 | −15.67% |
| Sep 2025 | −12.64% |
| Mar 2025 | −11.72% |
| Jul 2024 | −0.13% |
| Apr 2022 (low before 2024) | −6.18% |
How to read it
- A regression estimate, not a count of jobs: it compares the top and bottom fifths of occupations by AI exposure with firm-by-time fixed effects, which the AI Index says separate AI exposure from shocks such as interest rates or sector slowdowns.
- Values are percentage differences relative to November 2022, which is set to 0. Changes are in percentage points.
- Headcount by age for two exposed occupations is in the software developer chart.
Source: Brynjolfsson et al., 2025, via the AI Index Report 2026 (Stanford HAI), figure 4.4.30: Headcount trends in high AI-exposure jobs (early career 22–25), 2021–25. Chart and text by AI Adoption.
Numbers checked against the published data on 20 October 2026.
Download the data (CSV) · Cite this chart
Numbers checked against the published data on 20 October 2026.
Download the data (CSV) · Cite this chart
Discussion