OpenAI, Anthropic, Google in Talks to Form AI Standards Body

What Happened

OpenAI confirmed on Tuesday that it has been in discussions with rivals Anthropic and Google DeepMind for several weeks on artificial intelligence safety coordination, with talks focused on establishing a formal industry standards body. The confirmation, attributed to OpenAI policy chief Chris Lehane, was first reported by Bloomberg and subsequently corroborated by Reuters, CNBC, and TechCrunch.

OpenAI and Anthropic separately announced over the weekend that they would be slowing the development of certain frontier AI systems, according to the Jerusalem Post, a decision that coincided with the disclosure of the broader industry talks.

Background

The discussions trace back to at least July, when Google DeepMind chief executive Demis Hassabis published a proposal calling for a U.S.-led standards body to coordinate AI safety across major developers, according to a company spokesperson cited by CNBC. That proposal appears to have formed the basis for the current three-party conversations.

OpenAI, based in San Francisco, develops the ChatGPT family of models and is among the largest AI research organizations in the world by funding and deployed user base. Anthropic, also San Francisco-based, was founded in 2021 by former OpenAI researchers and develops the Claude family of models. Google DeepMind, the AI research division of Alphabet, is responsible for the Gemini model series and has a decades-long research track record spanning both academic and applied AI.

The three organizations collectively represent a significant share of frontier AI development globally, and any jointly adopted safety standards would carry substantial weight across the industry.

The Broader Context

The safety talks are taking place against a backdrop of escalating concern within parts of the AI research community. An Anthropic researcher's warning that advanced AI could pose extinction-level risks to humanity drew widespread attention in the days preceding the announcement, according to Neowin. A separate report from Ynetnews noted that a senior Google AI researcher, Bilal Chughtai, recently resigned and issued a public warning that the technology could, in his words, "kill us all."

At the same time, TechCrunch reported that the Trump administration has dismissed safety concerns and is pressing the industry to accelerate development in order to maintain pace with Chinese AI programs. That political dynamic places the three companies in a position of pursuing voluntary safety coordination without current federal regulatory backing in the United States.

France 24 reported that rising public concern over AI risks is one factor driving the push for a standards body, as the companies seek to establish shared benchmarks and evaluation frameworks before government mandates are imposed.

What It Means in Practice

Details of the proposed standards body remain limited in public disclosures. The talks are described as ongoing, and no formal founding documents, membership criteria, governance structures, or timelines have been published as of Tuesday. OpenAI's confirmation via its policy chief signals that the discussions have advanced to a stage where the company is prepared to acknowledge them publicly, but no signed agreements have been announced.

The initiative, if formalized, would represent the first jointly governed safety coordination structure among the three leading U.S.-based frontier AI developers. Earlier industry bodies, including the Partnership on AI, have included these organizations as members but operate on a broader, less binding basis.

Stocktwits noted that the announcement comes as the industry faces external pressure to slow frontier development, a pressure point that appears to have accelerated the timeline of the talks.

What Comes Next

The three companies have not announced a date for a formal agreement or public rollout of any standards body, and further disclosures are expected as negotiations continue.