OpenAI, Anthropic, Google reportedly building joint frontier AI safety authority

emmanuelvivier · x · 2026-09-25

OpenAI, Anthropic and Google are reportedly preparing their own "Standards Authority for Frontier AI" — an industry-run body for third-party model testing and incident reporting, targeted for early 2027 (unconfirmed). If true, the frontier labs would be building a self-regulatory mechanism outside government oversight.

Related event: Google, OpenAI and Anthropic Building Self-Regulated Frontier AI Standards Body SAFA(11 posts)→

Original post →

More from Safety

Safety channel →