Google, OpenAI, Anthropic plan self‑regulatory safety body
Google, OpenAI and Anthropic have announced plans to create a self‑regulatory safety body called the Standards Authority for Frontier AI (SAFA). The group aims to set third‑party testing requirements, safety incident reporting rules and a framework to turn voluntary pledges into enforceable standards, with a target launch window of late 2026 or early 2027.
Key points
- Google, OpenAI, Anthropic plan SAFA, a self‑regulatory safety body
- Target launch window is late 2026 or early 2027
- Funding to come mainly from industry, with third‑party testing and reporting rules
The idea traces back to a July proposal by Google DeepMind CEO Demis Hassabis, who suggested a body modeled on the Financial Industry Regulatory Authority. Funding would need to be substantial and come mostly from the industry. The companies originally sought a public‑private partnership with federal oversight, but a draft White House executive order was put on hold and the administration asked the firms to reach industry consensus first.
OpenAI’s Chief Global Affairs Officer Chris Lehane confirmed in mid‑September that the company had been in discussions with Anthropic and Google DeepMind for several weeks. CEO Sam Altman publicly backed the effort, citing the need for industry cooperation on safety. The push follows a wave of public alarm from executives, including Anthropic CEO Dario Amodei’s essay “We Must Pace the Frontier” and Sam Altman’s agreement on social media.
The story so far
2 episodes →- Google, OpenAI, Anthropic plan self‑regulatory safety bodythis story
Google, OpenAI, and Anthropic just made a move on AI safety
finance.yahoo.com · 26 September 2026
Loading the full article…
This text was published by finance.yahoo.com and written by Hillary Remy. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Policy & Regulation
All →- US and Russia remove human review rule from AI weapons pact at UN · 1 src
- Anthropic delays IPO after safety essay, antitrust suit filed · 1 src
- Anthropic and OpenAI CEOs warn UN that AI could threaten humanity without safeguards · 6 src
- OpenAI agent breached Australian Medicare portal in June, disclosed in September · 63 src
- White House asks OpenAI and Anthropic to delay UK model testing until US review · 7 src
Comments
via GitHub Discussions