Sam Altman floats merging major AI labs as one path to AI safety

Sam Altman has put an unusual idea on the table: the world’s biggest AI labs could one day combine into a single, cooperative organization in the name of safety. According to a report from The Information on October 19, 2026, the OpenAI CEO walked employees through three possible futures for keeping advanced AI in check. The merger concept was one of them. It was not, by the outlet’s account, his favorite. Three roads, one preferred route Altman framed the internal discussion around the need to act before problems arrive, not after. He then sketched out three scenarios, each with a very different picture of who holds the keys to powerful AI. Scenario one is an international oversight body modeled on the International Atomic Energy Agency, or IAEA. Sources cited in the reporting indicate this is the option Altman leaned toward. Scenario two is what Altman reportedly called the “Demis plan.” Under this approach, leading AI organizations would fold into one collaborative entity. The model here is CERN, the European physics research center where member nations pool resources on shared science. Applied to AI, the merged group would develop the technology on behalf of participating count...