Overview
- Multiple outlets reported on Sept. 24 that Anthropic, OpenAI and Google DeepMind are working to launch an independent, industry-led standards body by late 2026 or early 2027.
- The proposal would create shared protocols such as third-party model testing, pre-release safety checks, rules for reporting security incidents, and formal qualifications for independent auditors.
- OpenAI policy chief Chris Lehane has confirmed coordination among the three firms and said the collaboration would not require antitrust exemptions.
- No formal organization, charter or named leader has been announced, and reports that the body would be called Standards Frontier for AI or that Sriram Krishnan was approached to lead it remain unverified.
- Critics warn the initiative could entrench the largest labs, raise compliance costs that hurt smaller and open-source developers, and revive questions about funding, enforcement and international coordination.