{"type":"post","id":"ai-safety-nonprofit-metr-has-published-a-proposal-0b8b05","slug":"ai-safety-nonprofit-metr-has-published-a-proposal-0b8b05","title":"AI safety nonprofit METR has published a proposal for recurring assessments of frontier AI developers.","description":"METR’s proposal makes recurring external assessment a process for producing public evidence, rather than a pass/fail certification.","retrieval_nugget":"METR’s proposal makes recurring external assessment a process for producing public evidence, rather than a pass/fail certification.","published_at":"2026-10-15","updated_at":"2026-10-15","record_date":"2026-10-15","date_kind":"published_at","topics":["evals","governance","verification"],"source_urls":["https://metr.org/blog/2026-10-14-better-informing-the-public-about-loss-of-control-risks"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:ai-safety-nonprofit-metr-has-published-a-proposal-0b8b05","status":"published","visuals":[],"editorial_provenance":{"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:24cdf74d4fadc788b4a688ddf0ce7a96ef563d386b160e4f1cdfa4ad38f0500c","reviewed_at":"2026-10-15T04:04:59.203Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":2,"watch_condition_count":1,"related_record_count":0},"analysis":{"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Recurring external assessments can narrow the public evidence gap about frontier developers’ control of internal AI systems, but they cannot certify safety.","observed_facts":[{"text":"METR proposes weeks- or months-long assessments on company-managed devices, with access to internal systems and datasets and private staff interviews.","source_urls":["https://metr.org/blog/2026-10-14-better-informing-the-public-about-loss-of-control-risks"]},{"text":"METR would write public reports within an agreed scope; developers may choose access and request redactions.","source_urls":["https://metr.org/blog/2026-10-14-better-informing-the-public-about-loss-of-control-risks"]},{"text":"METR says the proposal does not require employee-level access, cannot fully protect against retaliation, and does not cover security audits.","source_urls":["https://metr.org/blog/2026-10-14-better-informing-the-public-about-loss-of-control-risks"]}],"mechanism":"Assessors combine internal evidence and confidential staff interviews, then publish a scoped report subject to developer access choices and redactions.","why_now":"METR points to recent incidents and growing calls for third-party evaluation, while saying public visibility into internal development remains inadequate.","implications":["Public reports could let outsiders compare evidence about alignment failures, monitoring and response, and AI-driven R&D acceleration across developers.","The usefulness of the process depends on scope disclosure and what companies permit assessors to inspect."],"evidence_boundary":"This is a proposal, not proof of broad developer participation or a finding that any developer is safe or unsafe; METR says the science is immature, access is discretionary, retaliation protections are incomplete, and security audits are outside scope.","watch_conditions":["Watch which developers undertake assessments, what access they grant, and whether reports disclose scope, limitations, and timely findings."],"related_records":[],"new_branch_reason":"No prior New Runtime record was supplied; this proposal starts a distinct record about METR’s recurring-assessment design and its limits."},"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/ai-safety-nonprofit-metr-has-published-a-proposal-0b8b05/","markdown":"https://newruntime.com/posts/ai-safety-nonprofit-metr-has-published-a-proposal-0b8b05.md","json":"https://newruntime.com/posts/ai-safety-nonprofit-metr-has-published-a-proposal-0b8b05.json"},"related_shifts":["verification-ownership"],"next_reads":[{"type":"topic","path":"/topics/evals/","url":"https://newruntime.com/topics/evals/","title":"Evals","reason":"Explore the evals topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/evals.md","json_url":"https://newruntime.com/topics/evals.json"},{"type":"topic","path":"/topics/governance/","url":"https://newruntime.com/topics/governance/","title":"Governance","reason":"Explore the governance topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/governance.md","json_url":"https://newruntime.com/topics/governance.json"},{"type":"topic","path":"/topics/verification/","url":"https://newruntime.com/topics/verification/","title":"Verification","reason":"Explore the verification topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/verification.md","json_url":"https://newruntime.com/topics/verification.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/source-ledger.json","url":"https://newruntime.com/source-ledger.json","title":"Source Ledger","reason":"Inspect the public source evidence graph.","media_type":"application/json"}]}