As part of the announcement of the 2023 Nobel Prizes, a scientist and his students have tested whether chatbots would be able to predict the winners. Will they be a tool used for this purpose in the future?
The announcement of the 2023 Nobel winners has driven an intriguing topic: Can artificial intelligence anticipate who the winners will be? This is a question that has been asked in the corridors of academic institutions.
Finally, the winners of the 2023 Nobel Prize in Physics are Pierre Agostini, Ferenc Krausz and Anne L’Huillier for their innovative tools to explore the universe of electrons within atoms, and no, they could not be predicted, even by artificial intelligence .
Before the announcement, scientist Santo Fortunato, from Indiana University Bloomington, United States, ventured to ask ChatGPT about its predictions, as reported Nature. However, at first the AI verdict was clear: can’t predict the future.
But heThe issue did not stop there.. Fortunato students asked a similar question to Claude, the Google chatbot. The goal was to identify the three major discoveries in chemistry, physics and medicine made by living scientists who have not yet received a Nobel Prize.
While the bots were able to identify significant achievements, such as CRISPR and the discovery of 2D materials like graphene, their answers were far from perfect. James Evans of the University of Chicago argues that language models like ChatGPT could become powerful prediction tools with the right modifications.
Global analytics company Clarivate has also begun exploring how generative AI could assist in predicting future Nobel laureates. In this scenario, AI would not replace human intuition and judgment, but could speed up and improve the identification of potential candidates.
Beyond the specific predictions, what is notable is the potential ability of AI to influence the selection process. James Evans of the University of Chicago sees a future where models like ChatGPT, with the right modifications, could be powerful prediction tools.
Artificial Intelligence Could Provide Deeper Insight into Potential Candidates
Nobel recognition is not only based on the number of published works or citations. As Rasmus Bjørk of the Technical University of Denmark points out, you need something “special”, work that really impacts the discipline and society. AIs, with their ability to analyze large amounts of data, could provide deeper insight into potential candidates.
However, not everything is rosy. There is a risk that AI will perpetuate historical biases if it is trained exclusively on data from previous winners. The conclusion is that, although the AI May Not Be Ready to Accurately Predict Nobel Prize Winnershas the potential to complement and possibly improve the selection process.