Anthropic gives outside evaluator METR staff-level access to its AI
The AI developer will let six independent researchers from the nonprofit METR examine its models, and METR can publish what it finds without Anthropic's approval.
By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .
This story is no longer being updated. It reflects reporting up to .
What happened
Anthropic has granted a six-person team from METR, a nonprofit AI evaluator, employee-level access to its systems, including company hardware, office space and internal tools. The access lets the group independently assess its advanced models, safety practices and risk management. METR will publish its own assessments, and Anthropic has given up the right to block them.
Why it matters
The arrangement gives outside evaluators a way to scrutinize unreleased models and publish findings the company cannot veto, which supporters see as greater accountability in AI safety.
Timeline
- Oct 15, 1:35 PM ET: Anthropic announced it had granted METR employee-level access to its systems.
- Oct 15, 4:35 PM ET: Anthropic also brought in the nonprofit SecureBio to help assess risks and safety practices, and is in early discussions with other organizations, including Hugging Face.
- Oct 16, 12:09 PM ET: Elon Musk said SpaceX would probably do the same.
Coverage
- Anthropic gives third-party auditors 'employee-like access' in AI industry first (See full bio)
- Six METR staff now work inside Anthropic, and it cannot veto what they publish (Shane S. Ellison)
- Anthropic will let outsiders see inside its company amid panic about rogue AI systems (the-independent.com)
- METR Gets Employee-Like Access Inside Anthropic, With Unconditional Right to Publish Findings (techtimes.com)
- Anthropic Gives METR 6 Staff AI System Access [2026] (tech-insider.org)
- Anthropic Grants METR Employee-Level Access to Assess AI Risks, Will Publish Results (NewsCord)
- Anthropic taps nonprofits METR and SecureBio to help assess risks and safety practices (CNBC)
- Anthropic is letting METR researchers look under the hood with employee-level access (techcrunch.com)
- Anthropic just opened its doors to outside safety watchdogs. Here's how it will work. (africa.businessinsider.com)
- Anthropic just opened its doors to outside safety evaluators (businessinsider.com)
- Anthropic grants employee-level access to nonprofit AI evaluator METR (seekingalpha.com)
- Anthropic grants AI evaluator METR employee-level access to its models (Reuters)
More from Technology & AI
- Eight states sue OpenAI, alleging misleading AI safety claims
- OpenAI announces solution to Navier-Stokes math problem
- Datadog bets AI agents can handle testing of AI-written code
- Anthropic and OpenAI sketch wealth-sharing plans before IPOs
- Mamdani administration proposes AI customer service rules
- Florida engineers study how AI behaves in high-risk situations