AI知识库 @ai521

most recent employer. Industry is assigned based on information reported by the employer and may be unknown for some customers. Race and Ethnicity: The customer’s race and ethnicity. Region: The region corresponds to the customer’s residence at the time of applying, regardless of where they were employed. California is divided into 14 regional planning units (RPUs) . How Are These Data Developed? (Methodology) This methodology, developed by CPL, involves linking California unemployment claims records to established measures of occupational AI exposure to track potential AI-related job loss over time. The data underlying the tracker are drawn from unemployment administrative records maintained by the EDD. These records are compiled as part of the operation of the program and cover workers who apply for benefits after a layoff from a covered California employer. Initial unemployment claim data reflect eligible workers who were laid off through no fault of their own, had sufficient prior earnings to qualify, and subsequently applied for benefits. Initial claims include both new claims (applied for the first time) and additional claims (when a customer returns to work, skips at least one certification week, and then reopens a prior claim before the benefit year expires). Each claim is dated to the month of its week-ending date. When applying for benefits, customers self-report their occupation from a standardized list. The tracker uses this occupation information to assign each customer to an AI exposure score based on their pre-layoff job. The data also include characteristics reported at the time of filing, including county of residence, industry of most recent employer (classified by the North American Industry Classification System, NAICS), age, gender, race and ethnicity, and educational attainment. AI Exposure Measures Each customer is assigned an AI exposure score based on the occupation they reported at the time of applying. The Tracker uses two complementary measures of AI exposure: Both measures are used together to build a more complete picture of AI exposure across the customer population. Results are largely consistent across both measures. Each customer is assigned to one of three AI exposure groups based on the score for their reported occupation: ● High AI Exposure : Top 25% of scores (potential measure: ≥ 0.49; observed measure: ≥ 0.107). Includes occupations most susceptible to AI-related disruption, such as customer service representatives and software developers. ● Moderate AI Exposure : Middle 25%–75% of scores (potential: 0.12-0.48; observed: 0.011–0.106). ● Low AI Exposure : Bottom 25% of scores (potential: < 0.12; observed: < 0.011). Includes occupations that are less susceptible to AI, such as heavy truck drivers or nursing assistants. Context for Using These Data ● AI exposure measures used in this tracker, including Anthropic’s Claude-based observed exposure measure, capture the extent to which job tasks could be or have been performed by AI, not whether AI was the cause of any specific layoff. The Tracker should be interpreted as a descriptive signal, not causal evidence, of AI-driven job loss. ● UI claims do not capture workers who do not file for UI, either because they are unaware of it, because they quickly find new employment, they exit the labor force altogether, or they are not eligible for UI due to their legal status, or because they are self-employed (including gig workers), or in the case of younger

7 views23:20