Google, OpenAI and Anthropic plan an industry safety standards body, the 'Standards Authority for Frontier AI', without government oversight
The Information reported on Sept 24, 2026 that Google, OpenAI and Anthropic plan to launch an independent industry body, tentatively the Standards Authority for Frontier AI (SAFA), in late 2026 or 2027. It would set safety standards, incident-reporting guidelines and auditor qualifications without government oversight. Talks on a federally supervised public-private body stalled after a draft White House executive order failed to get enough support.
Key facts
- Tentative name: Standards Authority for Frontier AI (SAFA); launch targeted for late 2026 or early 2027
- To be run by an independent CEO; candidates reportedly include former White House AI policy adviser Sriram Krishnan
- Planned functions: support third-party pre-deployment testing, guidelines for reporting safety and security incidents, rules for how voluntary lab commitments work, qualifications for independent auditors
- Origin: talks on a public-private partnership under federal supervision hit an impasse when a draft White House executive order lacked backing
What happened
The three leading US labs are working on a self-regulatory standards body after efforts to set one up under federal supervision stalled. It follows Hassabis's July proposal for a frontier AI standards body and OpenAI's Sept 21 call for US-led global technical standards.
Why it matters
With the White House rejecting binding oversight, the labs are moving to write common rules for testing, incident reporting and audits themselves. Critics are likely to question how independent such a body can be.
Caveat: based on anonymous sources in The Information (paywalled); name, timing and CEO candidates are tentative.
Changelog
- 2026-09-29: created (sweep 2026-09-29)
Related events
- Demis Hassabis proposes a US-led, FINRA-style Frontier AI Standards Body in essay "A Framework for Frontier AI and the Dawning of a New Age" ★★★★
- OpenAI calls for US-led global technical standards for frontier AI, including recursive self-improvement ★★★
- Altman and Amodei ask the UN Security Council for international AI standards and incident reporting ★★★★
- Anthropic and Accenture (Faculty) commit $1B+ to embedded third-party evaluation ★★★
Sources (4)
- pressThe Information: Google, OpenAI, Anthropic AI safety group takes shape
- pressThe Next Web: Google, OpenAI and Anthropic plan AI safety body
- pressTechRepublic: Google, OpenAI, Anthropic reportedly plan AI safety standards body
- pressPYMNTS: OpenAI, Google and Anthropic join forces to set AI safety standards
id: 2026-09-24-labs-plan-frontier-ai-standards-body-safa · updated 2026-09-29 · open in the interactive timeline