Amodei Calls for Slowing Frontier AI and Opening Anthropic to Independent Safety Reviewers

Anthropic CEO Dario Amodei published "We Must Pace the Frontier," calling for the strongest models to be developed more slowly and pledging permanent internal access for third-party safety evaluators; he also said the industry has long lied about AI risks. Altman and Musk backed his embedded-evaluator proposal, while safety researchers began challenging his arguments point by point.

Confirmed

• In a CBS interview (relayed by @imjustnewatai and @Polymarket), Amodei said the technology carries real dangers and that the industry has lied to the public about risks for too long.
• The essay's core claim: training can continue, but labs should stop pushing capability as fast as possible because safety work has not kept pace; it lays out a three-part plan (@eyishazyer, @tallmetommy).
• Amodei proposed that frontier labs accept "embedded evaluators" — independent safety staff inside companies verifying claimed training, deployment, operational and safeguards measures; per @MilesBrundage, Sam Altman and Elon Musk expressed support.
• Anthropic unilaterally committed to the first step: granting third-party evaluators permanent, employee-level access, including viewing how models are trained and conducting reviews without Anthropic arranging them.
• MIT Technology Review reported on the "doomer turn," noting a rare collective response from the four major labs (@nordicinst).
• Allie Miller (@alliekmiller) published a walkthrough unpacking the essay, sparking wide discussion.
• A safety researcher wrote a point-by-point rebuttal titled "Dario, please" (@GoMeansGo), showing Amodei's arguments face dissent even within the safety community.
• Per @SimplyAnnisa relaying eyishazyer, the safety debate is colliding with earnings season, a delayed IPO, industry coordination, and the repeatedly slipping Grok roadmap.

Unconfirmed

• The "OpenAI agent cluster cyberattack" cited by @nordicinst as background appears only in that account and awaits verification.

Why it matters

• A frontier-lab CEO publicly admitting industry dishonesty on risk communication and voluntarily granting internal access is a substantive governance concession, not mere rhetoric.
• With Altman and Musk endorsing embedded evaluators, adoption across labs could reshape the pace of the AI race and regulatory expectations.
• Point-by-point pushback from safety researchers shows disagreement persists within the safety camp itself.
• The debate's collision with earnings, IPOs and product roadmaps will test whether safety positions withstand commercial pressure.

2026-09-14 ~ 2026-09-16 · 9 related posts

Full story(20 episodes)→

• Episode 1: OpenAI employees speak out urging slower AI development over extinction risk(2026-09-10, 4 posts)
• Episode 2: OpenAI asks Congress whether coordinated AI slowdown would violate antitrust law(2026-09-11, 4 posts)
• Episode 3: Dario's "Pace the Frontier" Plea Draws Rare Backing from Altman and Musk(2026-09-11, 56 posts)
• Episode 4: Amodei, Altman and Musk unite on slowing AI, sparking debate over motives(2026-09-11, 62 posts)
• Episode 5: Anthropic CEO Dario Amodei Calls for Slowing Frontier AI Development(2026-09-12, 203 posts)
• Episode 6: METR Independence Under Fire: AI Audit Ecosystem Accused of Revolving Doors and Regulatory Capture(2026-09-12, 31 posts)
• Episode 7: Dario Amodei's "Pace the Frontier" Essay Draws Wave of Skepticism(2026-09-13, 30 posts)
• Episode 8: Amodei, Altman and Musk Rally Behind Slowing Frontier AI Development(2026-09-13, 20 posts)
• Episode 9: The Open Weights Regulation Debate: Tiered Controls, Ban Risks, and the Fate of Open-Source AI(2026-09-13, 18 posts)
• Episode 10: OpenAI researcher: taming superintelligence is science, not pacing deals(2026-09-13, 3 posts)
• Episode 11: 'China Will Slow Too If We Do' Claim Mocked as AI Race Meme Spreads(2026-09-13, 4 posts)
• Episode 12: Open-source camp pushes back on frontier slowdown pledges(2026-09-13, 4 posts)
• Episode 13: OpenAI researcher: labs' internal progress far exceeds outside perception(2026-09-13, 5 posts)
• Episode 14: Report: Anthropic and OpenAI May Be Slowing AI Research Ahead of IPOs(2026-09-13, 2 posts)
• Episode 15: François Chollet Questions Whether Frontier Labs' AI Slowdown Proposals Are Safety or Market Capture(2026-09-13, 6 posts)
• Episode 16: Coordinated AI Slowdowns Could Raise Antitrust Concerns(2026-09-13, 3 posts)
• Episode 17: Hassabis Backs Dario's AI Slowdown Call as Standards Bodies Weigh In(2026-09-13, 2 posts)
• Episode 18: Critics Call Out AI Labs' Safety Rhetoric as Hypocritical(2026-09-13, 3 posts)
• Episode 19: White House AI Chief Sacks Tells OpenAI and Anthropic to Slow Down Without Regulation(2026-09-13, 12 posts)
• Episode 20: Amodei Calls for Slowing Frontier AI and Opening Anthropic to Independent Safety Reviewers(2026-09-14, 9 posts)

• Anthropic CEO Dario Amodei: the industry 'lied for too long' about AI risks — Polymarket · 2026-09-14
• [source] Anthropic CEO Dario Amodei pledges independent safety reviewers, says industry "lied" about AI risks — imjustnewatai · 2026-09-14
• Dario Amodei Pens 'We Must Pace the Frontier,' Pledges Third-Party Safety Access — tallmetommy · 2026-09-15
• [source] Amodei says slow the frontier: Anthropic to grant third-party evaluators permanent staff-level access — eyishazyer · 2026-09-15
• MIT Tech Review: Amodei calls for LLM slowdown and top labs unusually agree — nordicinst · 2026-09-15
• [source] Dario Amodei's 'We Must Pace the Frontier' essay, explained — alliekmiller · 2026-09-15
• Amodei's safety debate collides with earnings season, IPO pushback, and slipping Grok roadmap — SimplyAnnisa · 2026-09-15
• Amodei proposes embedded safety evaluators; Altman and Musk endorse the plan — Miles_Brundage · 2026-09-15
• A security researcher's detailed rebuttal to Amodei's AI essay — GoMeansGo · 2026-09-16