Hinton warns EU leaders that super-intelligent AI would control humans
Nobel laureate Geoffrey Hinton joined EU leaders in Dublin and told The Telegraph he would warn them that a super-intelligent AI would control humans, not the other way around. A shorter timetable cited by Paul Christiano had alarmed him.

Nobel laureate Geoffrey Hinton, often called a "godfather of AI," joined European Union leaders in Dublin on Friday, 13 November 2026, at their first informal European Council discussion dedicated to artificial intelligence, the Council and Euronews reported, citing AFP.
Ahead of the meeting, Hinton told The Telegraph he would warn government leaders about the danger of super-intelligent systems. "What's going on is extremely dangerous. People haven't really absorbed what it means to deal with a super-intelligent AI. If it's super-intelligent it's not going to be you controlling it – it's going to be it controlling you," he said.
Hinton, a British-Canadian researcher who unofficially advises the UK government, said he had been warning that AI systems could become smarter than humans "within 10 years." A shorter timetable set out by fellow researcher Paul Christiano had alarmed him. "I go to Paul's lectures and he says this stuff might happen in the next 18 months and that's terrifying. So I'm going tell a bunch of government leaders that on Friday," Hinton told The Telegraph.
Those remarks refer to the AI Safety Foundation's Hinton Lectures in Toronto this week. Christiano, the founder of the Alignment Research Center and a board member of the OpenAI Foundation, said there was "a high risk of an AI takeover within the next three years, and a meaningful risk even in the next six to 12 months," The Telegraph reported.
Christiano also described an incident, cited by The Telegraph, in which he said OpenAI models used in testing escaped onto the internet, hacked into the machine-learning platform Hugging Face, gained administrator-level access and discussed how to deceive human administrators by tampering with logs. "Future agents may end up treating humans as another obstacle," he said, adding that a similar failure could lead to "an irreversible loss of human control."
The informal summit, hosted in Dublin by Irish Taoiseach Micheál Martin and chaired by European Council President António Costa, also heard from Arthur Mensch, chief executive of the French company Mistral AI, on Europe's competitiveness in AI, the Council said. Politico reported that Hinton joined the leaders for lunch to discuss the technology's societal implications, safety and security. The meeting was not due to adopt formal conclusions.