Anthropic launches $5M grant program for independent evaluations of AI's impact on user wellbeing
On Aug 25, 2026 Anthropic announced $5 million to fund independent, open-source evaluations of how AI models affect user wellbeing, especially in mental-health support and emotional companionship. Grantees keep full independence and receive funding, model access and technical support. Applications closed Sept 21, and full-proposal invitations are due Oct 5.
Key facts
- $5M total; open to clinicians, psychologists, methodologists and others
- Evals should define what they measure, involve domain experts, test both safeguards and harms, use realistic multi-turn conversations, and validate grading against experts
- Deadlines: applications Sept 21, 2026; notification of full-proposal applicants Oct 5, 2026
What happened
A grant call to build standard, open benchmarks for AI behaviour in sensitive emotional and mental-health conversations, run independently of the lab.
Why it matters
Mental-health harms from chatbots were a major 2025–26 policy and litigation issue. Shared, expert-validated wellbeing evals are a step toward measurable standards.
Changelog
- 2026-10-01: created (leads run, from the Anthropic uncited-posts audit)
Related events
Sources (1)
id: 2026-08-25-anthropic-wellbeing-evaluation-grants · updated 2026-10-01 · open in the interactive timeline