Anthropic Grants METR Employee-Like Access for Independent AI Safety Evaluations

Anthropic has started to give employee-like access to a team from METR, a nonprofit that evaluates frontier AI models. The company has onboarded six METR employees with company hardware, office space, and access to internal tools and systems roughly on par with its own risk assessment teams.

METR will publish its analysis of risk levels, incidents, and practices. Anthropic cannot veto publication, but can require redactions for narrow categories such as customer data, attorney-client privileged information, and model architectures. METR is permitted to publicly characterize whether it believes any redactions were justified without revealing the redacted information.

Anthropic is developing three tiers of access for independent evaluators. In the first tier, evaluators conduct narrowly scoped risk investigations and publish findings with Anthropic's approval. In the second tier, evaluators have access to more information and can publish without approval, subject to removal of specific information categories. The third tier provides employee-like access and publication rights.

SecureBio is currently in the first tier and is reviewing the chemical and biological weapons chapter of the August Risk Report, following a pilot review of a previous risk report. Anthropic is also working with Accenture and Faculty toward partnerships covering model evaluation, red-teaming, alignment assessments, safeguard testing, and capability building across organizations.