Sources
The AI Evaluator Forum's AEF-1 baseline for third-party evaluators now has xAI, OpenAI and Anthropic as cosigners
The 2026-09-15 AINews issue reports that AEF-1, a baseline standard from the AI Evaluator Forum (formed December 2025), has been cosigned by xAI, OpenAI and Anthropic, covering access, conflicts of interest, funding relationships, recusal and transparency. This moves Amodei's pacing proposal from a personal blog post into a signed multi-lab standard, with Anthropic committing to desks, access badges and company laptops for an embedded external review team at parity with internal risk-assessment teams. Medianama, Unite.AI and MarkTechPost independently report Altman matching the commitment, so the cosigning is corroborated beyond the aggregator.
↳ Follow the thread