California built a public tracker for AI job loss and so far it finds almost none

California's Employment Development Department and the California Policy Lab built a public tool that reads unemployment insurance claims against measures of how exposed an occupation's tasks are to AI, covering January 2019 through May 2026. The headline finding, according to the reporting, is a negative one: no evidence overall of rising claims from AI-exposed occupations. One subgroup moves — college-educated workers in high-exposure jobs, whose monthly claims rose from roughly 13,000 in late 2022 to between 16,000 and 22,000 since 2023. Researchers call these early signals, not a policy verdict, and caution that exposure captures what AI could do, not what employers actually adopted.