Hajra Naz
Chris Lehane, OpenAI’s global policy chief, spoke to reporters on Tuesday. He confirmed that OpenAI has been working with rivals Anthropic and Google DeepMind on AI safety. This collaboration has been ongoing for several weeks, as Bloomberg first reported. Lehane is reportedly in Washington right now. He’s working with U.S. lawmakers on legislation addressing catastrophic risks tied to AI.
This revelation follows an essay from Anthropic CEO Dario Amodei. He published it Saturday. In it, he called for industry-wide collaboration to slow the pace of frontier AI development. His goal was avoiding catastrophic risks. Several AI leaders across the industry supported his message.
That includes OpenAI’s Sam Altman. It includes Google’s Demis Hassabis. SpaceXAI’s Elon Musk also voiced support. Altman specifically said OpenAI would join Anthropic in embedding third-party evaluators. These evaluators would monitor for safety within the company.
On Tuesday, Lehane confirmed something specific. The three companies have been in talks for several weeks regarding AI safety coordination. Some people, including Altman himself, have raised a concern. Such coordination could risk violating antitrust law.
Read More: We Can Handle AI Safety Ourselves and We Don’t Need AI Regulation, Says Nvidia’s Jensen Huang
This applies if the collaboration is found to suppress competition unfairly. Amodei’s essay proposed a specific solution. He suggested a narrow government waiver, allowing this kind of safety coordination. However, Lehane reportedly said the companies don’t actually need one.
This news confirms earlier reporting and hints about the companies working together. In an interview with Fortune late last week, Altman hinted at something notable. He said he’d been in private discussions with other AI leaders.
This week, The Information reported additional details. The three companies have been working together on something specific. They’re creating a standards body for the broader AI industry. According to reports, Altman told staff this effort would need to happen. Notably, it would proceed without support from the U.S. government.
Hassabis called for something similar back in July. He urged the U.S. to establish a new standards body. This body would function as an AI watchdog of sorts. It would have the power to screen the world’s most advanced models. It could also coordinate industry-wide slowdowns if dangers progressed.
Read More: Microsoft’s New AI ‘Code of Conduct’ Says Models Shouldn’t Hack or Trick Humans
President Trump, for his part, has dismissed these safety concerns. He’s called them a “hoax.” He’s also pushed back against tighter regulations generally. He argues any slowdown would give China a competitive advantage in this race. His key AI advisor, David Sacks, echoed this sentiment. Sacks is a longtime investor with a personal stake in much of the industry. He said fears of existential risk from AI were overblown.
During that same Tuesday meeting, Lehane addressed another point. He said OpenAI supports a specific provision in the FRONTIER Act. This provision would require top frontier labs to allow “independent verification organizations” into their companies. The goal is ensuring models get developed safely.
