California’s new AI-Unemployment Tracker shows no statewide surge in unemployment insurance (UI) claims from highly AI-exposed occupations through May 2026, even as claims rose among college-educated workers, Bay Area claimants and some technology-sector groups.
The California Policy Lab developed the tracker in partnership with the California Employment Development Department. The Governor’s Office announced it as part of a state AI workforce initiative.
California officials described the dashboard as a “first-in-the-nation” tool to monitor possible AI-related job losses by linking unemployment insurance claims with occupational AI-exposure measures.
How the tracker measures AI exposure
The tracker groups claimants by how exposed their former occupation was to AI. One measure estimates whether large language models could reduce task time by at least 50%, based on work by OpenAI and academic researchers published in Science.
A second measure draws on the Anthropic Economic Index, which tracks how often tasks linked to occupations are performed using Claude.
A lack of statewide displacement
The first release found no evidence of rising statewide claims among AI-exposed occupations after the release of ChatGPT-3.5 in November 2022.
The report also found that the share of claims coming from AI-exposed workers did not increase in a statistically significant way compared with the pre-pandemic period.
Stress signals in the Bay Area and tech sectors
The statewide result raises other stress signals. Claims from college-educated workers in highly exposed occupations rose after ChatGPT-3.5’s release and stayed elevated through May 2026.
The California Policy Lab also reported a more than 50% increase in high-AI-exposure claims in the San Francisco Bay Area after ChatGPT-3.5’s release. Statewide, technology-heavy sectors such as Information and Professional Services also showed elevated claims, with Professional Services remaining higher relative to the broader state trend.
“Right now, we are not seeing evidence of large-scale AI-related layoffs in California’s labor market,” said Ben Hyman, senior researcher at the California Policy Lab and a report co-author. “But we do see patterns in certain regions like the Bay Area, in certain tech-heavy sectors, and among highly AI-exposed workers with college degrees.”
The tracker will be updated monthly to monitor how these patterns evolve over time.
Data limitations and post-pandemic dynamics
The report treats those patterns cautiously. Its limitations section states that AI-exposure scores show whether tasks could be or have been performed by AI, not whether AI caused displacement at a specific workplace.
Unemployment insurance data also misses workers who do not file claims, quickly find new jobs, leave the labor force, are self-employed or are not eligible for benefits.
It also says the initial rise in 2023 likely reflects broader post-pandemic labor-market dynamics, while the more persistent elevation among college-educated workers and tech sectors into 2025 and May 2026 is less easily explained by COVID-era factors.
Testing whether the tracker detects AI-linked layoffs
The report also tested whether the tracker can detect known AI-linked layoffs. In a validation analysis, the California Policy Lab examined six large employers that had publicly announced “AI-driven” mass layoffs between March 2024 and April 2025, then matched planned layoff dates from WARN filings with unemployment insurance claims.
The claims rose after the planned layoff dates, with the report estimating that roughly 58% of impacted employees filed for unemployment insurance benefits within 10 weeks after adjusting for normal turnover.
The authors said the exercise suggests the tracker can detect AI-driven job loss when it occurs, but those events have not reached a scale large enough to appear in aggregate statewide trends.
Echoing broader national labor trends
The California release adds a state administrative signal to other attempts to measure AI’s labor effect in live data. Stanford Digital Economy Lab and ADP Research’s Canaries Dashboard uses a five-year balanced sample of firms using ADP payroll services and reported slower employment growth in some AI-exposed occupations, especially among younger workers.
Stanford also cautions that the dashboard measures correlation, not causation, and does not represent the whole U.S. labor market.
The Budget Lab at Yale reached a similarly cautious national reading, saying in its June 15 tracker update that AI usage measures showed no connection to changes in employment or unemployment. The analysis also said the occupational mix was not yet changing in ways clearly aligned with AI’s introduction into the workforce.