Anthropic: Labor market impacts of AI – A new measure and early evidence [pdf] (cdn.sanity.io)

Anthropic has introduced a new framework to assess the labor market impacts of AI, focused on a metric called "observed exposure." This metric combines theoretical capabilities of large language models (LLMs) with actual usage data, emphasizing automated tasks over augmentative ones. The research highlights that while certain professions show high potential for AI automation, actual deployment lags significantly behind theoretical capabilities. Notably, workers in highly exposed occupations tend to be older, female, more educated, and higher-paid; however, there has been no substantial rise in unemployment among these groups since late 2022, although hiring for younger workers in these fields has slowed. The findings are significant as they seek to establish a more accurate understanding of how AI will reshape employment rather than relying solely on past forecasts, which have had mixed results. By equipping policymakers and businesses with a clearer picture of AI's potential impacts, Anthropic aims to better identify vulnerable jobs before significant displacement occurs. The work indicates that as AI technology progresses and its usage widens, tracking the gap between theoretical and practical applications could yield critical insights into future economic disruptions.