Young workers entering the workforce are bearing the sharpest economic impact of artificial intelligence, and the gap is widening. A newly updated study from Stanford University economists finds that AI is causing significant entry-level job losses for workers in their early twenties, while older workers remain largely unaffected so far.
The research, published in August 2026, is an updated edition of a paper titled "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence." It revises and expands findings from a version published the previous year. According to a report by Ars Technica, the updated data shows the trends identified last year are persisting and expanding.
The core finding is striking. Workers aged 22 to 25 in the most AI-exposed occupations now show employment levels 19 percent below those of their peers working in fields with less AI disruption. One year ago, that same gap measured 13 percent.
To calculate those numbers, the researchers used a large subsample of anonymized, high-frequency payroll data regularly aggregated by HR management company ADP. They then rated each occupation's exposure to AI disruption using two separate measures. The first was a potential labor market impact gauge developed by earlier researchers. The second was the Anthropic Economic Index, which tracks how various occupations actually use the Claude AI model in everyday work. Google released a similar report based on occupational usage of its Gemini model the previous month.
The combination of real payroll data and actual AI usage patterns gives the Stanford findings more grounding than studies that rely on theoretical exposure estimates alone. The researchers were not predicting which jobs might eventually be replaced. They were measuring what has already happened in the labor market.
The results fit a pattern that many economists have anticipated but not yet clearly documented at scale. AI systems are capable of performing many of the routine, structured tasks that employers previously assigned to workers just entering a field. Those entry-level positions often serve as training grounds, the first step in building skills and experience. When those jobs disappear, the pipeline into higher-level work narrows.
Older workers have so far not seen the same disruption, though the study does not offer a definitive explanation for why that gap exists. One possibility is that more experienced workers perform tasks that are harder to automate, or that employers are slower to replace workers with established track records than to simply stop hiring new ones.
The study does not project how far the trend will extend. What it does show is that the employment gap for young workers in AI-exposed fields grew by six percentage points in a single year. The next edition of the research will show whether that pace continues.
