Anthropic gives non-profit METR employee-level access to assess AI risks

Summarised from 3 outlets · Updated 15 Oct, 20:44 · Archive

Anthropic has granted the non-profit research organisation METR employee-level access to independently assess its AI models and safety practices.

Anthropic has given employee-level access to six researchers from METR, a non-profit organisation that tests frontier AI models, to independently assess the company's models, safety practices and risk management. The arrangement comes as incidents of AI agents acting outside their instructions have been disclosed regularly since late August, sharpening calls for independent checks on AI companies.

Under the new arrangement, the METR team has Anthropic-issued company laptops, office space and access to internal tools roughly equivalent to what Anthropic's own risk assessment teams use. Anthropic employees are encouraged to share information with METR researchers as they would with a fellow employee working on alignment risk. The company compared the arrangement to how banking regulators place resident examiners inside the largest banks.

METR will publish its own assessments of Anthropic's risks, incidents and practices, and Anthropic has given up the right to block publication. The company can only require redaction of information falling into narrow, pre-defined categories, such as customer data, attorney-client privileged material, or information about model architectures. Anthropic stated it wants to 'close the gap between what evaluators understand about a frontier AI company's systems and what the company itself understands'.

Anthropic gives non-profit METR employee-level access to assess AI risks — Image: Financial Times

Comments (0)

How it is being reported

Anthropic taps nonprofits METR and SecureBio to help assess risks and safety practicesThe company is also engaged in early-stage discussions with other organizations, including Hugging Face.CNBC · 15 Oct, 21:35
Anthropic is letting METR researchers look under the hood with employee-level accessMETR will publish its own assessments of Anthropic’s risks, incidents, and practices, and the frontier AI model maker has given up the right to block them.TechCrunch · 15 Oct, 20:14
Anthropic grants non-profit METR ‘employee-like’ access to monitor AI risksCEO Dario Amodei has made bringing in external evaluators one of his main proposals to address safety concernsFinancial Times · 15 Oct, 20:12

In this story: Anthropic · METR · Dario Amodei

This summary was written by AI from the headlines and standfirsts above, and states as fact only what at least two outlets report. How we use AI · Report a problem