Anthropic Taps METR and SecureBio for AI Risk Assessment

Anthropic, the AI safety-focused company behind the Claude family of models, has enlisted two nonprofit organizations — METR and SecureBio — to help assess its risk management practices and safety frameworks, according to a report from US Top News and Analysis. The move signals a push by the company to subject its internal safety evaluations to independent outside scrutiny at a moment when regulatory and public pressure on frontier AI developers is intensifying.

METR, which focuses on evaluating the capabilities and risks of advanced AI systems, and SecureBio, which specializes in biosecurity-related AI risks, bring distinct but complementary expertise to the partnership. By bringing in organizations with specialized domain knowledge, Anthropic appears to be targeting some of the most consequential threat categories associated with powerful AI — autonomous capabilities and potential misuse in biological contexts.

Read more Boeing 737 Max 10 Wins FAA Certification After Years of Delays →

The arrangement reflects a broader industry trend toward third-party auditing as AI companies face growing demands for transparency from lawmakers, researchers, and civil society groups. Independent evaluations are increasingly seen as a credibility marker that internal red-teaming alone cannot provide, particularly as model capabilities advance rapidly.

Anthropologic is also in early-stage conversations with other organizations about similar arrangements, including open-source AI platform Hugging Face, suggesting the company intends to expand the scope and diversity of its external safety assessment network. How formal those additional partnerships ultimately become remains unclear.

Continue reading at US Top News and Analysis