AI company Anthropic has provided employee-level access to six researchers from METR, a non-profit focused on testing frontier AI models. This arrangement, announced on Thursday, allows METR to independently evaluate Anthropic's models, safety practices, and risk management.
The METR team has been given Anthropic-issued company laptops, office space, and access to internal tools comparable to those used by Anthropic's own risk assessment teams. Anthropic has stated that its employees are encouraged to share information with METR researchers as they would with colleagues working on alignment risk.
Under the agreement, METR will publish its assessments of Anthropic's risks, incidents, and practices without the company's right to block publication. Anthropic can request redactions for specific categories, such as customer data or attorney-client privileged information, but METR is permitted to publicly describe these redactions and state whether it believes they were justified.
This initiative follows a September proposal by Anthropic CEO Dario Amodei to embed outside evaluators within AI labs. OpenAI has also pledged a similar approach, and both companies, alongside Google, Meta, Nvidia, and xAI, have signed a White House accord committing to independent review.