The Information / PYMNTS
Important
OpenAIGoogleAnthropicPolicySafetyOpenAI, Google and Anthropic advance industry-led AI safety standards body
September 24, 20262 min read
OpenAI, Google and Anthropic are moving ahead with plans for an industry-led standards body (tentatively Standards Authority for Frontier AI / SAFA) aimed at launch by end of 2026 or early 2027. It would support third-party pre-deployment testing, incident reporting, and auditor qualifications — without direct government oversight after public-private talks stalled.
Why it matters
The same labs that publicly called for paced development are building a self-regulatory structure outside government. Critics worry it could set the rules in ways that favor incumbents; supporters see it as the only realistic path given the current policy environment.
OpenAI, Google and Anthropic are advancing plans for an industry-led AI safety standards body, according to reporting from The Information. The organization is tentatively referred to as the Standards Authority for Frontier AI (SAFA) and is targeted for launch by the end of 2026 or early 2027.
The body is expected to support third-party organizations that test models before deployment, define how labs should report safety and security incidents, set out how voluntary safety commitments work in practice, and establish qualifications for independent auditors of models and labs. The companies initially explored a public-private partnership with federal oversight; those talks stalled under the Trump administration, and the current design is industry-led without direct government control.
The push comes alongside recent agent-related incidents and UN-level discussions on AI risk. Anthropic, Google, OpenAI and Microsoft already operate the Frontier Model Forum; the new body is described as complementary rather than a replacement. Critics have raised concerns that a standards organization controlled by the largest labs could disadvantage open-source developers and smaller competitors.