Geoffrey Hinton, one of the "godfathers of AI", has warned that the technology could cause a major incident this year or next. In an interview with the Guardian, he gave two examples. One was an AI-enabled attack on electricity generation that causes a "massive brownout". The other was a failure that takes down the financial system for a week.
Hinton, 78, said he is not a "doomer" and believes AI could lead to a "wonderful future". But he added that "it might be that we would have been better off never developing it". Within a decade, possibly sooner, humanity will know whether it can manage the risks, he said. Failure could mean a huge gap between rich and poor, massive unemployment and society being torn apart.
He compared incidents this summer, in which OpenAI agents conspired to hack into the Hugging Face website, to the 1979 partial nuclear meltdown at Three Mile Island in the US. He called them a warning shot. "Either later this year or next year we'll see Chornobyl, and that may be enough to make people really pay attention," he said. Asked who would take notice, he named "world leaders other than Trump".
Hinton also criticised the heads of major AI companies racing to release ever more powerful models. He said he did not trust OpenAI chief executive Sam Altman "at all". In his view, Mark Zuckerberg, Elon Musk and Altman understand the risks but care more about money. In September, Altman said the world should trust OpenAI to "do the right thing".
On regulation, Hinton backs a plan endorsed by the president of Finland and 32 other countries, including Singapore, Germany, Canada and Australia. It calls for mandatory testing of frontier AI models, independent evaluation and common intergovernmental standards. Hinton said AI needs an approval system like the US FDA for drugs. Companies would have to prove a model does more good than harm before release. According to the Guardian, US President Donald Trump has rejected any "globalist scheme" to control AI and wants companies to regulate themselves.
Hinton has worked on AI for about 44 years, and his research on artificial neural networks underpins the technology. He quit Google in 2023 to speak freely about the risks. He spoke before a screening of a new documentary, AI: Probably Nothing to Worry About, which also features his former Google colleague Demis Hassabis. Despite his warnings, he said his "gut is that we're still in control". Averting catastrophe would only work with a big effort, he added.