Hinton warns of AI-led blackout risk within a year or two
Geoffrey Hinton told The Guardian AI could trigger a financial or power shutdown within a year or two, criticised AI leaders and backed mandatory model testing.
The story in focus
Geoffrey Hinton told The Guardian that AI could trigger a large-scale incident, such as a temporary shutdown of financial or power systems, later this year or next. He criticised the risk-taking of AI company leaders and backed mandatory independent testing of frontier models before release, modelled on the FDA's drug approval process.
- Hinton said an AI-caused blackout of financial or power systems could come this year or next, describing it as a Chornobyl-style warning shot.
- He said he does not trust OpenAI chief Sam Altman at all and named Mark Zuckerberg and Elon Musk as leaders who understand the risks but prioritise money.
- He backs mandatory independent testing of frontier AI models, comparing it to the FDA's requirement for drug companies to show evidence of more good than harm.
- He said within a decade, possibly sooner, it will be clear whether the risks can be managed or society is torn apart.
- He remains on balance hopeful, saying he believes humans are still in control.
Criticism of AI leaders
Hinton said he did not trust Sam Altman, chief executive of OpenAI, at all. He named Mark Zuckerberg and Elon Musk alongside Altman as leaders who understand the risks but would rather have the money.
He called it extremely worrying that Altman is in charge of one of the frontier US AI companies racing China to build the most powerful systems.
Hinton quit Google in 2023 to speak freely about AI risks, saying he has a reputation that makes him a spokesperson and needs to burn it on this issue.
Testing proposal
Hinton backs a plan endorsed by the president of Finland and 32 other countries, including Singapore, Germany, Canada and Australia, for mandatory testing of frontier AI models, independent evaluation and intergovernmental common standards.
He compared it to the US Food and Drug Administration's drug approval process, in which companies must give an independent authority strong statistical evidence that a product does more good than harm before release. He cited the thalidomide scandal as showing that releasing first and suing later is the wrong approach, and said big tech companies want AI treated the way thalidomide was.
What incident does Hinton say could force action on AI risks?
He said a Chornobyl-style event could come later this year or next, such as an AI-enabled attack causing a massive brownout or an incident bringing down the financial system for a week.
Sources & further reading
Help keep the record accurate. Send a correction.
