OpenAI's global policy chief Chris Lehane confirmed on Tuesday that OpenAI has spent several weeks discussing AI safety with rivals Anthropic and Google DeepMind, according to reporting from Bloomberg. As reported by Bloomberg, Lehane told reporters the engagement with Anthropic and Google DeepMind has been underway for weeks, not days, during a briefing in Washington where he was meeting with U.S. lawmakers. The timing matters as it confirms earlier hints rather than announcing something new. Lehane, who is reportedly in Washington working with U.S. lawmakers on catastrophic AI risks, emphasized that the three companies do not need a government antitrust waiver to coordinate on safety matters. According to Reuters, Lehane said the companies did not see a need for a formal waiver to proceed, directly addressing concerns about potential regulatory scrutiny. Lehane argued that "it's better to try to work together to prioritize safety," as reported by Bloomberg. The discussions appear to focus on cooperation on AI safety rather than a broader agreement to slow or halt model development.
Polymarket traders assign roughly 25% odds to a frontier lab voluntarily pulling a model over safety concerns before the end of 2026, a figure that reflects trader positioning rather than any announced withdrawal. According to Polymarket data, the 25% 'Yes' price means the market still favors no pullback by year-end. Frontier AI refers to the most advanced, largest-scale models being developed by top labs, the systems capable enough that their failure modes carry outsized risk. That's the category at the center of this conversation, and it's why a prediction market has formed around the possibility of a lab hitting the brakes on its own release.
The talks follow an essay Anthropic CEO Dario Amodei published Saturday, calling on the industry to work together to pace frontier development and avoid catastrophic risks. As reported by TechCrunch, the essay ran to roughly 3,800 words, according to Times of India. Several prominent figures in the AI sector publicly backed Amodei's position, with OpenAI CEO Sam Altman, Google DeepMind CEO Demis Hassabis, and SpaceXAI's Elon Musk each expressing support. This comes as Altman says OpenAI would join Anthropic in embedding third-party evaluators to monitor model safety. The Information reported that the three firms have quietly discussed forming a shared AI safety standards body, though no agreement has been publicly announced. Hassabis called on the U.S. to establish a new standards body with authority to screen the world's most advanced models and coordinate industrywide slowdowns if dangers progressed. In a post on X, Altman said a slowdown had been a "primary topic" of discussion at OpenAI in recent weeks and that the company would have more to share "soon."
OpenAI is backing a key provision of the bipartisan FRONTIER Act that would require leading AI developers to allow independent third-party organizations to assess their models and safety practices, according to POLITICO. As reported by Reuters, Lehane also said OpenAI supports a provision in the proposed FRONTIER Act that would require top frontier labs to admit independent verification organizations, external groups tasked with assessing whether advanced models are being developed safely. The roots of the talks go back to July, when Google DeepMind chair Demis Hassabis outlined a vision on X for a U.S.-led oversight body — structured as either a public-private partnership or a self-regulatory organization answerable to the federal government, along the lines of the Financial Industry Regulatory Authority — that could vet cutting-edge AI models and trigger industrywide slowdowns when threats emerged, according to an OpenAI spokesperson told CNBC. Lehane also said OpenAI is prepared to back bipartisan legislation on the Hill targeting catastrophic AI risks, including a federal governance framework put forward by Republican Representative Jay Obernolte and Democrat Lori Trahan, as well as a Senate effort championed by Majority Leader John Thune alongside Republican Ted Cruz and Democrat Amy Klobuchar, according to Bloomberg.
The safety coordination comes amid significant regulatory tensions, as the Trump administration has largely dismissed concerns about existential AI risk, with President Trump arguing that safety concerns are a hoax and any slowdown would give China a competitive advantage. According to Reuters, this places the leading AI companies in an unusual position: pressing Congress for safety guardrails at the same moment the White House is prioritizing speed and competitive advantage abroad. The prospect of competitors coordinating has drawn scrutiny from antitrust regulators, as FTC Chair Andrew Ferguson, speaking Tuesday, said he would be "deeply suspicious" of requests for antitrust exemptions, and that he views the AI companies as seeking barriers to entry that insulate their businesses from challengers, according to Bloomberg. "That sure sounds like moat digging," Ferguson said. The coordination is particularly significant given that the three organizations compete directly in the market for large language models and AI services. Reuters separately noted that the Federal Trade Commission has expressed suspicion toward calls for AI antitrust exemptions, a signal that regulators are watching how these agreements are structured. The debate has also raised concerns among investors about what a slower pace of AI development could mean for future spending on chips and data centers. Any AI bill becoming law before the end of 2026 is a long shot, as the House departs Washington on Thursday and has no scheduled sessions before voters head to the polls in November, according to Bloomberg.
