Anthropic grants METR researchers employee-level access to independently assess its AI models and safety

Anthropic provided six researchers from METR, a nonprofit testing frontier AI models, with employee-level access, including company laptops, office space, and internal tools, to independently assess Anthropic’s models, safety practices, and risk management. METR will publish its assessments without Anthropic's veto power, though Anthropic can require redactions for narrow, pre-defined categories like customer data or attorney-client privileged information. This top-tier access is part of a three-tier evaluation system Anthropic is building, fulfilling a promise made by CEO Dario Amodei in September.