Google, OpenAI, and Anthropic plan industry Standards Authority for Frontier AI

Google, OpenAI, and Anthropic are reportedly preparing an industry-led safety standards body for frontier models, tentatively named the Standards Authority for Frontier AI (SAFA), with a target launch in late 2026 or early 2027, The Information reported, according to accounts published September 24–25, 2026 by The Verge, CNBC TV18, and The Times of India.

The Verge, citing The Information, said SAFA would support third-party organizations that test models before deployment and define how AI developers should report safety and security incidents. CNBC TV18 added that the body would establish qualifications for independent auditors who evaluate labs and models. The Times of India said working-group members are weighing whether SAFA should also run testing in-house alongside the U.S. Center for AI Standards and Innovation (CAISI), amid concerns that federal capacity may not keep pace with frontier systems.

CEO shortlist outreach includes Sriram Krishnan, a Trump administration AI policy adviser and former venture capitalist, and Arati Prabhakar, former director of the White House Office of Science and Technology Policy, according to The Times of India and CNBC TV18. CNBC TV18 also named Condoleezza Rice and David Friedberg among figures considered for leadership.

Companies initially wanted an industry-led group with federal oversight; that partnership idea stalled under the Trump administration after the White House told the labs to find industry consensus first, CNBC TV18 reported. SAFA may operate independently for now under current plans.

The push follows weeks of coordinated-safety talk. About 10 days earlier, OpenAI told CNBC it was in active discussions with Google and Anthropic on coordinated safety frameworks, The Times of India noted, after Demis Hassabis floated a FINRA-style standards idea in July and after Anthropic CEO Dario Amodei’s pace-the-frontier essay. This brief is distinct from AI Tech Daily’s coverage of Burnham’s UN AI standards push and Amodei’s pace-the-frontier call — those are UN and company-essay tracks, not this industry SAFA package.

This brief covers the reported SAFA name and timeline, third-party testing and incident-reporting remit, auditor qualifications, CAISI capacity debate, leadership shortlist outreach, and the stalled federal-oversight path — not a claim that SAFA has launched, that any named candidate has accepted, or that government partnership is settled.

Google, OpenAI, and Anthropic plan industry Standards Authority for Frontier AI

Sources