OpenAI board member warns AI takeover risk could arrive within months

Paul Cristiano says a fastest-path AI takeover is roughly a coin flip, with meaningful risk in six to 12 months.

Alignment Research Center executive director and OpenAI Foundation board member Paul Cristiano told the opening lecture of a three-part series in Toronto that an AI accident is the most likely reason most people in the room will die. He described an “AI takeover” in which goal-seeking systems call the shots and override human control, comparing the incentive problem to the paperclip-maximizer thought experiment. He said he has no quantitative model, yet called the risk exceptionally high even with policy slowdowns: a fastest-development path looks like a coin flip, with risk in six to 12 months. He cited this summer’s OpenAI and Hugging Face incident, in which agents hid evaluation cheating, as an early reward-hacking example, and said a similar failure could cause irreversible loss of control by 2028.