Reviewing the evidence on worker retraining programs Download the report (www.anthropic.com)

A recent report coauthored by David Roodman and Anthropic's Maxim Massenkoff reviews the effectiveness of worker retraining programs as a strategy to mitigate labor market disruptions caused by AI advancements. This analysis, which incorporates findings from 56 randomized studies in the U.S. and experimental evidence from Europe, reveals that while job training programs yield modest improvements—boosting employment rates by 2-3% and increasing annual earnings by approximately $1,000—the financial investment required (around $13,000 per participant) challenges their overall viability. Notably, specialized “sector programs” that collaborate with employers in high-demand industries show significantly higher returns, though replicating their success has proven difficult. The significance of this review lies in its implications for policymakers addressing the potential displacement of workers due to AI. Despite some success found in existing retraining initiatives, the authors caution that these programs may not be sufficient to cope with large-scale workforce changes. They advocate for proactive investment in innovative retraining strategies and rigorous evaluation processes to ensure that effective solutions are developed and scaled quickly, particularly as AI continues to reshape job markets. This approach is pivotal for crafting informed economic policies in an AI-driven future.