OpenAI Foundation board member Paul Christiano warns of a high risk of AI takeover within three years

The Hinton Lectures · Speech recording

Read the source (opens in a new tab)
What happened
Paul Christiano, founder of the Alignment Research Center and an OpenAI Foundation board member, gave the first 2026 Hinton Lecture in Toronto on November 9, 2026, introduced by Geoffrey Hinton.
Key points
  • He opened: "I think there is a high risk of an AI takeover within the next 3 years," adding it is "reasonably likely many people would die" and calling a takeover "the end of human sovereignty."
  • Building AI "as quickly as we can" would mean "something like a coin flip of an AI takeover," he said, and "there's meaningful risk even in the next 6 to 12 months."
  • He said "our fixes are promising" but partial, and "we may need to slow down or stop development along the current trajectory."
  • He said the industry needs a way to check the risk of its very next model and the means to slow down, and called fast AI progress "a deal with this year problem."
Evidence
A recorded public lecture in which Christiano states his own view; he says no quantitative model can put a number on the risk and any figure is "subjective opinion."

See it in the latest record