When your robot wins a Nobel Prize

We’re in the midst of the Nobel Prize season, and once again, all the awards have gone to humans. But this doesn’t have to be the case forever. At the rate artificial intelligence (AI) is developing, it may not be long before a robot deserves science’s most prestigious awards. What will the Swedish Academy do then? Give it to them? Why not?

The idea, in fact, is nine years old. In 2016, the Japanese scientist Hiroaki Kitano, director of the Institute for Systems Biology and chief technology officer at Sony, presented the Nobel Turing Challenge to the scientific community, which consists of developing an AI system capable of making a discovery worthy of the Nobel Prize. Kitano is convinced that one of the greatest obstacles to scientific progress is the limited nature of human cognition. The millennium competition is underway.

When he conceived the challenge, Kitano was thinking of something like 2050 as the year to reach it. But that was nine years ago, before ChatGPT took the world by storm. The large language models (LLMs) that underlie that digital chatterbox and dozens of other related systems have surprised their own creators and indicated that machines’ cognitive abilities are developing faster than expected.

One of the challenge’s organizers, chemical engineer Ross King of Cambridge University, now believes it’s possible we’ll see a robot win the Nobel Prize within 10 years. Perhaps we should start thinking about how to dress it for the ceremony in Stockholm. Will it look good in a morning coat or a long dress? Perhaps I should commission an AI fashion designer to design the suit? It’s all a mystery.

Machines have long contributed to aspects of scientific activity such as data analysis and experimental design, but none of this would be enough to win a Nobel Prize, of course. A sector that is often considered more specifically human is hypothesis generation, and this is where the debate becomes more interesting.

A Nobel Prize tends to be awarded for discoveries that are useful in some way, that have ramifications in other areas of knowledge and that prove fruitful, that is, that open up previously unexplored and even unsuspected paths to science.

AI is already helping with tasks we consider specific to the human mind: formulating hypotheses about the origin of the universe, predicting astronomical phenomena, designing quantum computers, or deciphering animal language. An AI called Coscientist conceives complex systems of chemical reactions at least as well as the best human experts, and certainly much faster.

Last year’s Nobel Prizes in Physics and Chemistry went to the scientists who conceived artificial neural networks and AI systems that have revolutionized the deduction of protein shapes from their sequence alone. Of course, the Swedish Academy didn’t reward these machines, but rather their creators, but there’s no problem in principle with machines designing other machines, systems generating other systems, and so on, until the human contribution is nothing more than a distant memory. Will the machine then deserve the Nobel Prize? Why not?

The company Sakana AI is developing systems to automate research in machine-learning systems. The models that result from this process will not actually be the creations of a human scientist. Chatbots like ChatGPT are beginning to be used to simulate the discussions and seminars in which scientists discuss their ideas. Computational biologists at Stanford University in California presented their Virtual Lab a year and a half ago for this purpose, and the Shanghai Artificial Intelligence Laboratory has done something similar with its VirSci system.

The idea that a robot might deserve a Nobel Prize in the coming years isn’t absurd. Whether it’s actually awarded is another matter entirely. We humans are extraordinarily squeamish when it comes to recognizing any merit in machines. Chess seemed like a feat of the human mind until May 11, 1997, when Deep Blue beat Garry Kasparov. Well, we said then, it beat him by brute force, and after all, winning at chess doesn’t necessarily imply being intelligent, does it? Since then, we’ve been constantly looking for excuses to preserve the chauvinism of our species. Remember that, when your robot wins a Nobel Prize.

This article is part of Tendencias (Trends), a project by EL PAÍS that aims to open a conversation about the major future challenges facing our society. The initiative is sponsored by Abertis, Enagás, EY, Iberdrola, Iberia, Mapfre, Novartis, the Organization of Ibero-American States (OEI), Redeia, Santander, WPP Media, and strategic partner Oliver Wyman.

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition