In 2023 he left Google so he could speak freely about AI dangers. He had warned that if the systems get carried away with getting more control, we are in trouble.
The computer scientist told the Guardian the shock could resemble Chernobyl in its effect on public attention, though not as a nuclear accident. He cited examples such as an AI-enabled attack that causes a massive brownout of electricity generation, or an incident that takes down the financial system for a week.
His role as a progenitor of modern AI has drawn comparisons with Robert Oppenheimer, who was central to the creation of nuclear weapons. A friend even bought him an Oppenheimer-style hat.
He contrasted that with what he described as big tech’s approach, saying they treat the technology as thalidomide was treated.
He has worked on AI for about 44 years. In the 1970s he helped coin the idea of the neural network, a mathematical system that can learn and gain experience in a way analogous to people.
Hinton, 78, said people could recover from such an event and that it might make governments and companies sit up and take notice. Geoffrey Hinton, a Nobel laureate often called one of the godfathers of artificial intelligence, said humanity might have been better off never developing the technology, warning it could trigger a large-scale incident this year or next.
Within a decade, and possibly sooner, he said, humanity will know whether it can manage the risks and enable what he called radical abundance, or whether it fails to manage those risks.
He is not a doomer about the field built on his research into artificial neural networks. He said AI could still lead to a wonderful future, but added it might be that we would have been better off never developing it. Speaking to documentary makers, he said he felt he had a duty to warn.