Anthropic has granted the nonprofit METR inside access to audit its AI models and safety practices. Six METR staffers received company hardware, office space, and access to internal tools, placing them roughly on par with the firm’s own risk teams. Customer data, information covered by attorney-client privilege, and details about the model architecture remain off-limits. METR will publish its findings independently. Anthropic can ask for redactions in certain predefined categories but METR is free to state publicly whether it thinks those redactions were justified.
This arrangement represents a significant shift in how the company approaches external oversight. Anthropic argues that not all safety-relevant information can be made public, either because it could help attackers or because it is commercially sensitive. The METR deal goes further than what the company has offered other outside reviewers. SecureBio, Accenture, and Faculty have so far received less access and tighter limits on what they can publish. Anthropic has started early talks with Hugging Face. The move follows Anthropic’s signing of the White House Accord on Super Intelligence, which includes a commitment to working with outside reviewers.
* METR staff operate with privileges similar to internal risk teams
* Redactions are permitted only in predefined categories
* Previous partners received restricted access and publishing limits
