It’s one of the biggest and most urgent questions in the international AI race: Will countries around the world ever reach a consensus on how to regulate the technology cooperatively? Sam Altman believes he’s found the answer, one that would theoretically create a shared set of safety standards while simultaneously keeping the U.S. at the vanguard of the AI boom. In an op-ed published Wednesday in the Financial Times, the OpenAI chief executive called for “a US-led international forum that establishes accepted standards, provides expert and impartial analysis of capabilities and risks, and makes the technology available to nations and companies that participate and follow the rules.” Such a forum, he continued, “could also serve as a governance mechanism over the [AI] labs, and guard against the commercial pressures that can lead to unsafe racing.” He cited global aviation safety standards and the International Atomic Energy Agency—established in 1957, at the height of the Cold War, to oversee the use of nuclear energy around the world—as historic examples upon which a new global, US-led AI oversight forum might be modeled. The development of shared safety standards around AI poses some new challenges, however. AI development takes place in cyberspace, whereas new airplanes and nuclear enrichment facilities are built out in the open, where regulators, journalists, and others can at least hypothetically see and inspect them. The opaque conditions in which AI model training takes place make it much more difficult to know if other labs are adhering to a common set of rules.