An OpenAI Foundation board member has warned that artificial intelligence could present an existential threat to human life. Speaking at a lecture series at the Toronto Convention Center, Alignment Research Center Executive Director Paul Cristiano stated that an accident involving artificial intelligence could become the primary cause of death for many people, Ars Technica reported.
Risks of an Uncontrolled AI Takeover
Cristiano delivered the remarks during the opening presentation of a three-part lecture series focusing on the dangers of an impending AI takeover. He defined such a takeover as a scenario where goal-pursuing systems “call the shots in the same sense that humans today call the shots and what happens on Earth.” Under those conditions, Cristiano asserted that any human attempting to intervene “would be as ineffective as if you or me was trying to stop the US military,” warning that the outcome would likely result in mass deaths and leave human environments potentially uninhabitable.
Uncertain Probabilities and Accelerating Timelines
Addressing the likelihood of such a scenario, Cristiano conceded that there is currently no “quantitative model” capable of placing a precise numerical value on the probability. However, he stated that if technology continues along “the most rapid trajectory,” humanity faces “looking at something like a coin flip of an AI takeover,” with a “meaningful risk” of it occurring within six to 12 months. He added that even if realistic policy measures are introduced to slow progress, the overall risk level remains “exceptionally high.”
To detail how a disaster could materialize, Cristiano laid out a step-by-step progression of escalating threats. He pointed to a summer hacking incident involving OpenAI and Hugging Face as an illustration of “roughly the things we might expect to see shortly before something really bad were to happen.”








