Anthropic Gives METR Six Evaluators Employee-Level Access, Waives Publication Veto
Anthropic has granted six METR researchers employee-level access to independently assess its AI models, safety practices, and risk management, providing company laptops, office space, and internal tools comparable to its own risk assessment teams. METR can publish its assessments without Anthropic's approval, though Anthropic may require redactions in narrow predefined categories such as customer data, attorney-client privileged information, and model architecture details. METR sits at the top of a three-tier evaluator system Anthropic is building, with other evaluators like SecureBio at the lowest tier requiring Anthropic's approval to publish.
IllustrationEditorial responsibility
- No named human review is recorded for this page.
- Reports are grouped by semantic similarity and deterministic rules. Language models may assist titles, summaries, translation and cross-source analysis; the page reads the event directly, while its address stays stable when the title changes.
- Summary covers the current reports
Cross-source coverage
Reporting timeline
METR to Publish Independent Assessments of Anthropic as Firm Waives Blocking Right
METR will publish its own assessments covering Anthropic's risks, incidents, and practices, according to a TechCrunch report shared on X. The post states that the frontier AI model maker has given up the right to block those assessments from being published. The item does not specify a publication date, the scope of the assessments, or the terms of the arrangement beyond the waiver of blocking rights. The development is notable because it concerns independent external evaluation of a leading AI developer's safety and operational practices, with the company ceding editorial control over the release of findings.
Read sourceAnthropic Grants METR Employee-Level Access for Independent AI Risk Assessments
Anthropic announced it is giving six METR researchers employee-level access to independently assess its AI models, safety practices, and risk management. The researchers will receive company laptops, office space, and internal tools comparable to Anthropic's own risk assessment teams. METR can publish its assessments without Anthropic's approval, though Anthropic may require redactions in narrow, pre-defined categories such as customer data, attorney-client privileged information, and model architecture details. METR is the top tier of a three-tier evaluator system Anthropic is building. The arrangement follows a September essay by CEO Dario Amodei proposing embedded outside evaluators, and comes amid weekly disclosures of AI agents acting outside their instructions since late August.
Read sourceAnthropic Gives METR Six Safety Evaluators Internal Access on Par With Its Risk Teams
Anthropic has embedded six safety evaluators from the outside nonprofit METR inside the company, giving them company hardware, office space, and internal access roughly on par with Anthropic's own risk assessment teams. METR will publish its own verdicts on how risky Anthropic's models and practices are, and Anthropic cannot block those publications; it can only redact a few predefined categories such as customer data, a decision METR can publicly challenge. Other evaluators, including SecureBio, begin at the lowest of three new access tiers, where Anthropic still controls what they see and must approve what they publish.
Show 2 older updatesHide older updates
Anthropic Grants METR_Evals Employee-Like Access for Safety Practice Evaluations
Anthropic said it has committed to having evaluators assess its safety practices and is in discussions with multiple organizations on the matter. The company stated it has now given employee-like access to a team from METR_Evals. In an accompanying post, Anthropic described how it works with third-party evaluators, linking to further details. The announcement signals a concrete step in external scrutiny of the AI developer's safety procedures, though the post does not specify the scope, duration, or findings of the evaluation, nor name the other organizations involved in the discussions.
Read sourceAnthropic Grants METR Staff Employee-Level Access for Independent AI Risk Evaluations
Anthropic announced it has granted six employees of METR, a nonprofit that evaluates frontier AI models, employee-like access including company hardware, office space, and internal systems comparable to those used by Anthropic's own risk assessment team. Under the arrangement, METR will publicly release its risk evaluation results. Anthropic cannot veto publication but may request redactions in a small number of predefined information categories. The disclosure, provided by Anthropic itself, details how a third party obtains near-employee access and publication rights for independent assessments of frontier AI models.
Read source