AI winning Nobel prize: when that could come?

AI can revolutionize research and create breakthroughs

Artificial Intelligence (AI) is already profoundly impacting a wide range of industries, including banking, finance, the film industry and journalism. Researchers are also exploring how AI can significantly change research and even create Nobel Prize-worthy breakthroughs.

In 2021, Japanese scientist Hiroaki Kitano proposed the 'Nobel Turing Challenge', which aims to create an 'AI scientist' capable of conducting Nobel Prize-level research independently by 2050. Scientists are already actively pursuing this goal by trying to build an AI that can generate significant scientific discoveries independently.

According to Ross King, professor of machine intelligence at Sweden's Chalmers University, there are currently around 100 so-called "robot scientists". He has contributed to the development of the field himself and in 2009, together with colleagues, introduced 'Adam the Robot Scientist', the first AI device capable of making scientific discoveries on its own.

This year's Nobel Prize winners will be announced on 7-14 October, and scientists will continue exploring how AI could revolutionize their fields.

"We have built a robot that has made new scientific discoveries on its own, generated new scientific ideas, tested them and confirmed that they are correct," said King.

The robot is designed to generate hypotheses and design experiments to test them. He also programmed the lab robots to carry out those experiments, learning from the process and repeating it.

Adam was able to discover previously unknown functions of the body's genes - but he's no scientist

Adam was tasked with studying the inner workings of yeast and discovered the previously unknown functions of the organism's genes. Later, a second robot scientist, Eve, was created to research new drugs for malaria and other tropical diseases. According to King, robot scientists have several advantages over average human scientists.

They cost less money, they work 24/7, and robots also record every detail of the process more meticulously. However, King admitted that AI is not the equivalent of a Nobel Prize-winning scientist - they would need to be much smarter and able to understand the bigger picture.

The scientific profession is not in danger shortly

Inga Strumke, Associate Professor at the Norwegian University of Science and Technology, said that the scientific profession is not in a dangerous situation shortly because machines cannot yet take over from science. She stressed that although this is not yet possible today, there is no denying that artificial intelligence is already having a significant impact on science and this impact will only grow in the future.

One example of how AI is already contributing to scientific research is AlphaFold, an AI model developed by Google DeepMind that predicts the three-dimensional structure of proteins based on their amino acid sequences. This greatly facilitates the study of proteins and opens up new possibilities in medicine and biology.

"We knew that there was some relationship between amino acids and the final three-dimensional shape of proteins... and then we could use machine learning to find it," said Strumke.

Inga Strumke explained that the calculations performed by AI models such as AlphaFold are too complex for human capabilities. However, she said, the AlphaFold case also highlights one of the weaknesses of current AI models: while it is good at processing vast amounts of information and providing solutions, it cannot yet adequately explain why those answers are correct.

"AlphaFold's ability to predict more than 200 million protein structures is very useful, but, as Strumke says, these results "have not given us insight into microbiology". This means that while AI models provide enormous benefits, they still have a limited understanding of the very biological systems they study.

In the next decade, AI-assisted research will win Nobel Prizes

According to Inga Strumke, the scientific goal is to understand the universe, not just to make correct predictions. Nevertheless, the achievements of AlphaFold are so significant that the scientists who developed the technology are considered to be serious contenders for the Nobel Prize.

"Google DeepMind's director John Jumper and the company's CEO and co-founder Demis Hassabis have already been awarded the prestigious Lasker Prize in 2023 for their pioneering work. In addition, the Clarivate think tank, which tracks potential Nobel Prize winners in science, considers both of them as likely candidates for the 2024 Chemistry Prize, the winner of which will be announced on 9 October.

David Pendlebury, head of the research team, notes that while their 2021 paper on AlphaFold has received thousands of citations, the Nobel Prize committee usually awards discoveries decades after they were made. Nevertheless, he is confident that AI-assisted research will win Nobel Prizes in the next decade.

Based on ELTA reports