Geoffrey Hinton warned that an AI-enabled attack on a power grid could cause a major brownout, while another incident could disable the financial system for a week, exposing risks to critical infrastructure.
Hinton said he believed a major incident could occur this year or next, The Guardian reported Tuesday. He described a grid attack that could produce a massive brownout, as well as a separate event that could take the financial system offline for a week.
Hinton said AI could generate what he called “radical abundance,” but also warned of unemployment, widening inequality and social disruption, The Guardian reported. He questioned whether developing AI had been the right choice, but told the newspaper he did not consider himself a “doomer.”
He linked the dangers partly to major AI companies’ competition for technological supremacy and market share, according to the interview. Hinton compared recent incidents involving AI agents with the 1979 partial meltdown at Three Mile Island, The Guardian reported.
At Starmus VIII in Tenerife on Oct. 20, Hinton urged developers to make AI systems retain concern for people even if they could eventually rewrite their own code. He also criticized technology companies for focusing too heavily on capability and too little on safety, as EconoTimes reported in its coverage of Hinton’s argument that AI systems should retain concern for people. He compared the relationship between people and more capable AI to a parent caring for a dependent child.
The Royal Swedish Academy of Sciences awarded Hinton and John J. Hopfield the 2024 Nobel Prize in Physics for work that enabled machine learning with artificial neural networks.
Hinton said people may know within a decade, possibly sooner, whether AI’s risks can be managed, The Guardian reported.
