Amodei Proposes Embedded METR Oversight of Frontier AI, Sparking Independence Backlash
Anthropic CEO Dario Amodei published "We Must Pace the Frontier," proposing that nonprofit evaluator METR (Model Evaluation and Threat Research) be embedded in frontier AI companies with employee-like access to oversee safety commitments—analogous to bank examiners resident in banks. Embedded third parties would have the power to report safety incidents, assess model alignment, and publish honest conclusions, with Anthropic pledging access for METR, Redwood Research and others. Sam Altman endorsed independent inspectors for both Anthropic and OpenAI the next day, but the arrangement triggered a broad backlash over METR's independence.
Confirmed
- Amodei's essay argued for third-party evaluation teams embedded like employees at frontier labs (@firstadopter, @basedjensen); NY Post ran a critical report on the safety proposal.
- Sam Altman backed independent inspectors at Anthropic and OpenAI; one thread noted the context of Anthropic researcher Jacob Coxon's resignation and public statements (@alexverem).
- METR self-assessed its independence against its own standardized criteria, performing well on most items while lacking any disclosed conflict-of-interest policy, buried mid-way through a 300-plus-page report (@kevinnbass).
- @kevinnbass traced the previously contested Coefficient funding through intermediary layers, claiming it ultimately points to Anthropic investors (@kevinnbass, @firstadopter).
- Gary Marcus argued METR's world overlaps with Anthropic's, so it cannot be the sole third-party auditor, while acknowledging its research quality (@GaryMarcus).
- Very Sane AI Newsletter published a long response arguing METR cannot check Anthropic's power (@kyliebytes).
- @segyges clarified he was not asserting METR is necessarily conflicted, but raised concerns about mechanisms like labs recommending their own auditors; the post was amplified by jdpressman as a significant take (@inductionheads repost, @jessicata).
Unconfirmed
- The funding chain allegedly ending at Anthropic investors is the poster's own inference, not confirmed by METR or the investors involved.
Why it matters
- The dispute cuts to the core of frontier AI governance: embedded third-party evaluation loses its checks-and-balances value if the auditor's interests intersect with the audited company; METR's self-assessment conspicuously lacking a conflict-of-interest policy amplifies concerns about the independence of the evaluation ecosystem.
2026-09-15 ~ 2026-09-17 · 12 related posts
- Episode 1: Open Philanthropy Accused of Fueling AI Doom Narrative With Over $1 Billion(2026-09-12, 8 posts)
- Episode 2: Anthropic Researcher's Doomsday Claims Alleged as Orchestrated PR Campaign(2026-09-14, 5 posts)
- Episode 3: Hugging Face co-founder sparks debate on AI labs' profit motives and danger claims(2026-09-14, 7 posts)
- Episode 4: Debate Rages Over Whether AI Doom Talk Is Regulatory Capture(2026-09-15, 4 posts)
- Episode 5: Blogger's "Audit" Claims Anthropic Built a Regulatory Capture Machine, METR Independence Disputed Then Rebutted(2026-09-15, 16 posts)
- Episode 6: Amodei Proposes Embedded METR Oversight of Frontier AI, Sparking Independence Backlash(2026-09-15, 12 posts)
- Episode 7: Researcher Answers Dario's Slowdown Call: Responsibility, Not Rationing(2026-09-16, 3 posts)
- Episode 8: Ben Todd blasts OpenAI and Anthropic over "pacing the frontier"(2026-09-16, 2 posts)
- Episode 9: METR's Independence Questioned: Client Capture and Worldview Alignment(2026-09-16, 7 posts)
- Episode 10: METR's $71M Six-Month Funding Surge Draws Transparency Scrutiny(2026-09-17, 9 posts)
- Episode 11: Anthropic's Dario Amodei Calls for Pacing AI Frontier; Musk, Altman and Hassabis Rally(2026-09-17, 6 posts)
Primary sources
- [source] METR rates its own independence highly, but omits conflict-of-interest disclosure policy — kevinnbass · 2026-09-15
- METR self-audit passes most of its own criteria, fails only on conflict-of-interest disclosure — kevinnbass · 2026-09-15
- Gary Marcus: METR's Ties to Anthropic Mean It Can't Be the Sole Auditor — GaryMarcus · 2026-09-15
- [source] Amodei proposes bank-style embedded regulators for AI labs; Altman signs on same day — alex_verem · 2026-09-16
- Debate erupts over whether METR can be impartial in auditing Anthropic's models — jessi_cata · 2026-09-16
- AI eval drama: critic argues METR cannot be impartial about Anthropic — inductionheads · 2026-09-16
- Essay argues METR cannot meaningfully check Anthropic's pace commitments — kyliebytes · 2026-09-16
- NY Post blasts Anthropic's safety proposal as Amodei pushes employee-level access for METR — basedjensen · 2026-09-16
- METR's independence questioned: EA ties and Anthropic investor funding fuel conflict-of-interest criticism — firstadopter · 2026-09-16
- [source] Amodei's METR embedded-evaluator proposal draws fire: "Not the SBF crowd" — firstadopter · 2026-09-16
- Anthropic and OpenAI pledge to embed third-party safety evaluators, but true independence remains an open question — RebeccaBellan · 2026-09-17
1 near-duplicate retellings: kevinnbass