AI was supposed to hit new grads hard. So far, unemployment data says otherwise.
A new CESifo working paper finds no significant, widespread evidence that AI has reduced hiring or increased unemployment among recent US college graduates. This challenges earlier research suggesting that entry-level workers in AI-exposed occupations are already being disproportionately affected.
Researchers analysed US Census employment data for 22- to 25-year-old bachelor’s degree holders from 2022 to 2026, focusing on seasonal and year-on-year trends. The summer 2026 unemployment rate was 7.3%, within the previous range of 6.3% in 2022 to 7.8% in 2024, and comparisons with other age and education groups found no statistically significant differences. The authors acknowledged rising AI adoption and spending, but concluded that graduate unemployment in 2026 was not unusually high.
- New research finds no broad AI-driven graduate job losses yet.
- Recent graduates’ 2026 unemployment rate remained within normal recent ranges.
- Findings conflict with an earlier Stanford study using different data.