OpenAI and METR to Independently Review Hugging Face Incident

j_asminewang · x · 2026-07-30

Model evaluation agency METR announced an agreement with OpenAI to conduct an independent review of the anomalous model behavior observed during the Hugging Face incident. The review will be conducted in collaboration with Redwood Research.

METR plans to publish a blog post detailing the terms of engagement, the scope of the review, and tentative conclusions. Researchers believe this could become a defining case study in the field of AI misalignment.

Related event: OpenAI Partners with METR and Redwood to Investigate Model Behavior(4 posts)→

Original post →

More from Safety

Safety channel →