Google, OpenAI and Anthropic plan joint AI safety standards body

By Splitfeed EditorialLast updated Sep 26, 2026, 19:17 UTC

On September 24, 2026, reports emerged that Google, OpenAI, and Anthropic are working together to establish a joint AI safety standards body. The organization has a tentative name, Standards Authority for Frontier AI, and is expected to launch either by the end of 2026 or in early 2027. The announcement came alongside the same three companies releasing or discussing new frontier models, with Anthropic drawing separate notice from Bessemer for what was described as consistency in its AI safety commitments.

The timing is significant because it places a cooperative safety initiative directly alongside an ongoing debate about whether leading AI labs are moving too quickly. Critics and observers have framed this tension as the "pace the frontier" problem: companies simultaneously setting the pace of development and proposing to govern it. Whether an industry-led body can produce genuinely independent standards, or whether it will function more as a credibility shield, is the central question the announcement raises. The choice to make it a standalone organization rather than embed it within an existing regulatory structure suggests the companies want flexibility, but that same flexibility is likely to be contested.

None of the three camps, Pro-AI, Anti-AI, or Middle Ground, had published specific reactions at the time of writing. Pro-AI voices would typically treat this as evidence that the industry can self-regulate responsibly and that voluntary coordination among frontier labs is a faster and more technically informed path to safe outcomes than government mandates. Anti-AI voices would typically argue that companies defining their own safety standards creates a conflict of interest, and that launching new models while simultaneously announcing a safety body does little to slow the risks being introduced. Middle Ground observers would likely call the initiative a useful first step while pressing for details about enforcement mechanisms, third-party auditing, and what consequences, if any, would follow a violation of the standards.

The clearest thing to watch is whether the Standards Authority for Frontier AI actually launches within the stated window and, if it does, what its governance structure looks like. Specifics around membership criteria, whether companies outside the founding three can join, and whether any government bodies have an oversight role will determine whether the organization is treated as a credible check or dismissed as an industry talking point.