OpenAI board member warns AI takeover risk could materialize within months

Paul Christiano, who serves on the OpenAI Foundation board and its Safety and Security Committee, told the audience at the Hinton Lectures 2026 that an artificial-intelligence accident could become the most likely cause of death for people working in the field. He warned that a meaningful risk of an AI takeover exists within the next six to twelve months and that, if AI labs continue to develop technology without intervention, the odds of a takeover could resemble a coin flip within three years.

"Even with realistic policy responses, the risk remains exceptionally high," Christiano said. Christiano previously estimated a 4% chance of catastrophic loss of human control within one year and a 15% chance within three years, noting these figures are subjective judgments rather than model-derived predictions. He added that OpenAI is unlikely to develop AI at its maximum speed, which he said reduces the risk to a level comparable to a dice roll rather than a coin flip. Other AI safety researchers have expressed similar concerns. Eli Lifland of the AI Futures Project tweeted that rapid self-improvement in AI could leave humanity without control within three to twelve months. Buck Shlegeris, CEO of Redwood Research, suggested a roughly 20% chance that AI misalignment could cause severe societal disruption or loss of life within a year. The remarks highlight ongoing debate within the AI community about the balance between rapid development and safety oversight, as well as the role of expert warnings in influencing policy and funding decisions.