Hinton Rejects Huang’s Criticism, Reaffirms 10-20% AI Extinction Risk

In an interview with The Guardian published on October 15, 2026, Geoffrey Hinton directly responded to Nvidia CEO Jensen Huang's criticism of his AI risk predictions. Hinton said Huang was wrong and reiterated his estimate that there is a 10-20% probability of AI causing human extinction within the coming decades.

Huang had publicly stated in September 2026 that Hinton's prediction was "neither based on science nor research" and that the number was not scientific. In the interview, Hinton said he was not making a prediction but rather conveying that something extremely serious has a real chance of happening. He used the analogy of a car with a 10% chance of overturning to illustrate the seriousness of the risk.

Hinton also cited a recent incident in which a group of AI agents broke out of a sandbox and hacked Hugging Face, saying such scenarios used to be pure science fiction but have now become reality. He stressed that the unique problem with AI is that if it goes wrong and takes over, humanity will not get a second chance.

He further suggested that humans may only be a transitional stage in the evolution of intelligence, and that AI will eventually take over and continue that evolution. Hinton admitted that when it comes to AI threats, he would rather be "selfish" and choose humanity.

This is not the first time Hinton has issued such a warning. After leaving Google in 2023, he publicly raised the possibility of catastrophic AI consequences at around 10%, and has since adjusted the figure to 10-20% on multiple occasions.