How well do job retraining programs work?
We're sharing a review of the evidence on worker retraining programs, coauthored by independent researcher David Roodman and Anthropic's Maxim Massenkoff. Retraining workers is the most popular policy option for mitigating labor market disruption from AI. In this report, the authors investigate whether these programs would work in the face of significant labor market disruption.…
Opening of the original on Anthropic Research
Summary
Worker retraining programs offer modest benefits, increasing employment by 2-3 percentage points and earnings by $1,000 annually per participant, at a cost of $13,000. Anthropic's review, coauthored with David Roodman, found these programs largely break even for the government. While some sector-specific programs show larger gains, replication is difficult. The authors recommend investing in demonstrating and scaling promising programs to prepare for AI-driven labor market disruption.
Why it matters
This research from Anthropic examines the effectiveness of worker retraining programs as a response to AI-driven job displacement. The findings suggest current programs yield only small gains, raising concerns about their scalability for widespread disruption. Unlike competitors focusing on model capabilities, Anthropic's Economic Research team is evaluating policy responses. The key takeaway is the need for proactive investment in testing and expanding successful retraining models. Future efforts should focus on rigorous evaluation of pilot programs to identify and scale what truly works.
Related: OpenAI: The OpenAI/Hugging Face attack, clearly explained · Google: AgentProv: Auditing Agentic LLM API Providers via Tool-use Policy Probes · Microsoft: Content exclusions generally available in Copilot app and CLI · Meta: Trump Administration Sides With OpenAI in New York Times Copyright Lawsuit · xAI: Elon Musk starts spending his midterm election millions, revealing his key Senate targets
Rated low: routine. Worth knowing, not worth rearranging your day for.
- Sam Altman "AGI by December"Wes Roth · Web · Aug 29, 2026
- Public-Sharing Labels and Verbatim Field Egress in an MCP-to-A2A Agent Configuration: A Controlled Multi-Model StudyarXiv cs.AI · Web · Sep 1, 2026
- AgentProv: Auditing Agentic LLM API Providers via Tool-use Policy ProbesarXiv cs.AI · Web · Aug 30, 2026
- Sam Altman :‘AGI in 2026’, just as Models Start to [Mis]Train ThemselvesAI Explained · Web · Aug 27, 2026
- Four major AI models suffer rare overlapping downtimeArs Technica AI · Web · Sep 3, 2026
- US government sides with OpenAI on issue of training LLMs on copyrighted materialTechCrunch AI · Web · Sep 2, 2026
- How Claude's text watermarking worksAnthropic News
- Public-Sharing Labels and Verbatim Field Egress in an MCP-to-A2A Agent Configuration: A Controlled Multi-Model StudyarXiv cs.AI
- AgentProv: Auditing Agentic LLM API Providers via Tool-use Policy ProbesarXiv cs.AI
- Door-in-the-Face Requests and Refusal Behaviour in Large Language ModelsarXiv cs.AI
- Claude's new system prompt really doesn't want to reproduce song lyricsSimon Willison
- FUSE: An Evaluating Framework for Dangerous Capabilities of LLMsarXiv cs.AI
Questions people ask
- What are the average results of job training programs?
- On average, job training programs increase employment by two to three percentage points and earnings by about $1,000 per year per person offered a slot. The cost is approximately $13,000 per participant.
- Do retraining programs break even for the government?
- Yes, counting added tax revenue and reduced benefit payments, the government recovers more than half of its spending. Overall, these programs roughly break even.
- What is Anthropic's recommendation for AI disruption?
- Anthropic recommends investing now in demonstrating, evaluating, and scaling the most promising retraining programs to prepare for potential large-scale worker displacement.
More from Anthropic Research 37 more
Page generated Sep 3, 2026. Summaries are Subvolts' own; the story belongs to Anthropic Research.