It is a common refrain that artificial intelligence is coming to take all our jobs, and now it seems that Nobel prizewinners are no exception. Two of the awards this year, for physics and chemistry, have been claimed by people working in the field of AI 鈥� much to the chagrin of some researchers in areas more traditionally recognised by these categories. What does the rise of the AI Nobel mean for the future of science?
鈥淭hese prizes reflect two different ways of reckoning with the relationship between AI and science: as a tool for studying the world and as a worthy pursuit in its own right,鈥� says at the University of Cambridge. 鈥淚n other words, they broadly correspond with the 鈥榮cience of AI鈥� and the 鈥榰se of AI in science鈥�.鈥�
The 2024 Nobel prize for physics falls into the first category, being awarded to Geoffrey Hinton at the University of Toronto, Canada, often dubbed one of the 鈥済odfathers of AI鈥�, and John Hopfield at Princeton University. Both were given the award for their work on machine learning 鈥� a key tenet behind AI.
Advertisement
Meanwhile, the chemistry Nobel has gone in part to Demis Hassabis and John Jumper, both at Google DeepMind, for their work on the AlphaFold2 AI model. This tool, which can predict the structure of 200 million proteins, is being used by researchers to better understand things like antibiotic resistance.
With AI mania sweeping the world since the release of ChatGPT in 2022, you could argue that the Nobel Foundation has been caught up in the hype. Some researchers wait decades to receive recognition for their work, but AlphaFold2 is just three years old, perhaps making the award premature.
鈥淚 would have preferred to see if it has had more direct ties to societal-changing protein or drug discoveries,鈥� says at the Georgia Institute of Technology, though he thinks the physics award was an entirely reasonable way to recognise the rapid pace of the field. 鈥淎I has had a huge impact on society in the last few years,鈥� he says. It is also worth noting that at the University of Washington, in Seattle, who received the other part of the chemistry prize, was rewarded for his work designing new proteins, which is unrelated to AI.
But as AI evolves and becomes more powerful, it is likely that AI-powered Nobels could become commonplace. In a statement responding to his win, Hassabis said: 鈥淚 hope we鈥檒l look back on AlphaFold as the first proof point of AI鈥檚 incredible potential to accelerate scientific discovery.鈥�
鈥淲e鈥檙e likely to see more AI-related research winning in the future,鈥� says at the University of Oxford. Drug discovery and other labour-intensive areas that can be quickly outsourced to AI could be ripe for future prizes, for example.
And with AI companies hoping to develop machines with human-level intelligence, could an AI model itself ever be awarded a Nobel prize? Hassabis, speaking at a press conference following his win, says such talk is 鈥渇ar too premature鈥�, but didn鈥檛 outright say no. V茅liz, however, doesn鈥檛 think so. 鈥淎I, despite the computer science lingo, is not an agent, and in particular, it鈥檚 not a moral agent,鈥� she says. 鈥淚t is not responsible for what it creates.鈥�
Instead of awarding a Nobel prize to an AI, Vliz would like to see more awarded to women. 鈥淚t should be noted that so far [this year] there have been seven male Nobel laureates and no women,鈥� she says. 鈥淲hat does that say about the Nobel prize?鈥�
