Geoffrey Hinton, widely called the godfather of artificial intelligence, has said humanity might have been better off if AI had never been invented. He is 78, has worked on the field for about 44 years, and holds a Nobel prize for the work that made modern AI possible.
The remarks came in an interview with The Guardian.
The specific thing he is worried about
Hinton's warning is not abstract. He expects that uncontrolled use of AI could produce a major catastrophe within this year or next, and he names the shape it would take: AI-enabled cyber-attacks on electricity supply, or attacks capable of paralysing financial systems.
He reached for Chernobyl as the analogy — suggesting that an incident of that kind might be what finally makes people take the problem seriously. That is a bleak thing for a founder of a field to say about it: that the warning will not land until something breaks.
He is not predicting doom
This is the part that gets lost when the quote travels. Hinton is explicit that he is not entirely pessimistic. AI could, he says, create genuine new prosperity — his condition is that the risks be managed alongside the benefits rather than after them.
He expects the answer within the next decade, possibly sooner. If the risks are managed, he describes large-scale abundance. If they are not, he describes the gap between rich and poor widening, widespread unemployment, a cultural crisis and the possible collapse of social structures.
And he says control is still in human hands, and that effective steps remain possible — requiring a sustained effort he believes humanity is capable of making.
The part that will make headlines
Hinton named names. He said he does not trust OpenAI's chief executive Sam Altman. He said Mark Zuckerberg, Elon Musk and Altman are all aware of AI's risks but, in his words, "don't care" — believing that making money matters more to them than human safety.
His wider concern is structural rather than personal: that American firms racing China may lack responsible leadership at exactly the moment it matters, releasing more powerful models one after another while showing insufficient interest in managing what they have released.
Why this particular person saying it matters
It is worth being clear about why Hinton's view carries weight, because "AI pioneer warns about AI" has become a genre.
Hinton began working on artificial neural networks in the 1970s, when the approach was unfashionable and largely dismissed. Much of his career was spent at the University of Toronto. The methods he persisted with are the direct ancestors of the systems now being deployed at scale.
In 2023 he left Google specifically so he could speak about the risks without representing an employer. That decision cost him something and removed the usual reason to discount what he says. He has no product to sell and no funding round to raise.
None of which makes him right. But it does mean his warning cannot be dismissed as unfamiliarity with the technology, and his timeline — this year or next — is short enough to be tested rather than argued about.
