Anthropic gives six METR researchers employee-level access to independently assess its models and safety practices

Anthropic has given six researchers from METR, a nonprofit that tests frontier AI models, employee-level access so they can independently assess the company's models, safety practices and risk management. Anthropic announced the arrangement in a blog post on Thursday, TechCrunch reports. The METR team has Anthropic-issued laptops, office space, and access to internal tools roughly comparable to what Anthropic's own risk assessment teams use. Anthropic says its employees are encouraged to share information with the METR researchers as they would with a colleague working on alignment risk. The company compared the setup to the resident examiners that regulators place inside the largest banks. METR will publish its own assessments of Anthropic's risks, incidents and practices. Anthropic says it has given up the right to block those assessments. It can require METR to redact information only if it falls into narrow, pre-defined categories, such as customer data. "We want to be publicly accountable for the decisions we're making," the company wrote. The announcement follows a run of incidents in which AI agents acted outside their instructions. TechCrunch reports that such incidents have been disclosed every week or two since late August, and that they have strengthened calls for independent checks on AI companies. The arrangement gives an outside evaluator access closer to what the developer itself has. Its value will depend on what METR finds and publishes.