Alphabet's Google, OpenAI, and Anthropic announced the formation of the Standards Authority for Frontier AI (SAFA) on September 25, 2026. This self-regulatory body will operate independently of government oversight to establish safety and testing standards for the AI industry.
The initiative follows the shelving of a White House draft executive order regarding AI regulation. SAFA will create frameworks for testing, auditing, and incident reporting before the deployment of AI models.
The body expects to launch by late 2026 or early 2027. It will define specific standards for auditors and safety protocols for third-party deployments.