Anthropic CEO proposes 'pace the frontier' plan with embedded third-party auditors

According to TechCrunch and The Verge, Anthropic CEO Dario Amodei proposed a plan to slow AI development by giving third-party evaluators like METR access to Anthropic's models to monitor adherence to safety practices. Amodei and OpenAI's Sam Altman agreed the field should decelerate development pace.

Update (2026-09-16): Anthropic CEO Dario Amodei published an essay calling for coordinated slowing of AI development with third-party auditors embedded in labs, citing risks from LLM progress outpacing safety capabilities, according to MIT Technology Review. OpenAI and Elon Musk have expressed support for the position, while Trump dismissed ...