World
Nobel physics prize awarded to ‘godfather of AI’ who warned the technology could end humanity
No surprise AI innovators awarded
Given the step-wise, often collaborative nature of scientific breakthroughs, it’s often hard to predict who might win the ultimate prize in science – a Nobel – and when.
But given the unbelievably powerful AI models unleashed on the world last year, you probably wouldn’t need GPT4 to predict a prize for the “inventors” of modern artificial intelligence was on the cards.
American physicist John Hopfield and British computer Scientist Geoffery Hinton pioneered the concept of neural networks – digital analogues of how nerve cells in the brain “learn”.
Receiving the news by telephone from Stockholm, Hinton, the so-called “Godfather of AI”, said: “I’m flabbergasted. I had no idea this would happen. I am very surprised.”
Though legendary in the world of AI and computer science for decades, Hinton only came to prominence last year when he warned of the existential dangers of powerful generative AIs that emerged in early 2023.
He has since spoken of his regret in laying the foundations for that technology.
Asked about his work in front of the Nobel Prize committee, he said, “I’d do the same again, but I am worried what the overall consequence of this might be. Systems more intelligent than us that eventually take control.”
This Nobel is interesting in that some other key players in the development of neural networks aren’t included, like Yan LeCun, now chief AI scientist at Meta, and Yoshua Bengio, of the University of Montreal.
The logic of the Nobel committee is perhaps that this is the prize for physics, not computer science, and Hopfield and Hinton’s breakthroughs depended most on application of statistical physics.
Reacting to the news of his prize, Hinton credited his former supervisor, the psychologist David Rumelhart, as the real pioneer of AI.
It was Rumelhart who supervised Hinton’s work on neural networks and came up with the “backpropagation algorithm” – the key concept that underpins today’s large generative AI models like GPT4, Gemini and Bard.
Rumelhart died in 2011 however, and Nobel Prizes are only ever awarded to living scientific pioneers.