How the AI ​​Nobel Prizes could change the focus of research

But Hodgkinson fears researchers in the field will look more to the technique than the science when trying to figure out why the trio won the prize this year. “I hope this doesn’t lead to researchers using chatbots inappropriately, mistakenly thinking that all AI tools are equivalent,” he says.

The fear that this could happen is based on the explosion of interest in other supposedly transformative technologies. “There are always hype cycles, the most recent being blockchain and graphene,” says Hodgkinson. After graphene was discovered in 2004, 45,000 scientific papers mentioning the material were published between 2005 and 2009, according to Google Scholar. But after Andre Geim and Konstantin Novoselov won the Nobel Prize for their discovery of the material, the number of published papers skyrocketed, to 454,000 between 2010 and 2014 and to more than a million between 2015 and 2020. This surge in research has arguably been modest so far Real World Impact.

Hodgkinson believes that the stimulating power of several researchers recognized by the Nobel Prize jury for their work in AI could lead to others joining forces in the field – leading to science of variable quality. “Whether the proposals and applications (of AI) have substance is another question,” he says.

We have already seen the impact that media and public attention on AI is having on the academic community. According to a Stanford University study, the number of AI-related publications tripled between 2010 and 2022, with nearly a quarter of a million articles published in 2022 alone: ​​more than 660 new publications per day. That was before the November 2022 release of ChatGPT kick-started the generative AI revolution.

The extent to which academics are likely to follow the media attention, money and praise of the Nobel Committee is a question that concerns Julian Togelius, an associate professor of computer science at New York University's Tandon School of Engineering who studies AI. “Scientists generally pursue a combination of the path of least resistance and the best value for money,” he says. And given the competitive nature of science, where funding is becoming increasingly scarce and is directly linked to researchers' job prospects, it seems likely that combining a trending topic – as of this week – with the potential to give high achievers a Nobel Prize might be too tempting to resist.

The risk is that this could hinder innovative new thinking. “It is difficult to obtain more fundamental data from nature and develop new theories that are understandable to humans,” says Togelius. But that requires deep thinking. Instead, it is far more productive for researchers to run AI-enabled simulations that support existing theories and incorporate existing data – resulting in small advances in understanding rather than giant leaps. Togelius predicts that a new generation of scientists will do just that because it's easier.

There is also a danger that overconfident computer scientists who have helped advance the field of AI will realize that AI work in unrelated scientific fields – in this case physics and chemistry – is being awarded Nobel Prizes and decide to move into theirs To follow in footsteps and encroach further on other people's turf. “Computer scientists have a well-deserved reputation for poking their noses into areas they know nothing about, throwing in some algorithms and calling it progress, for better or for worse,” says Togelius, who admits to having previously tried to add something Taking deep learning to another area of ​​science and “taking it further” before he changes his mind because he doesn’t know much about physics, biology or geology.