2026 World Laureates Forum Opens: Nobel and Fields Medalists Discuss How AI Is Changing the Scientific Research Paradigm

The deep integration of artificial intelligence and scientific research is brewing a profound transformation. Faced with the entirely new possibilities brought by AI, one Nobel laureate is exploring a new way of working, while a Fields Medalist returns to the source of mathematical creation, asking how mathematicians should understand a "machine that can prove." On October 31, the 2026 World Laureates Forum opened in Shanghai under the theme "Paradigm Shift: The Next Decade of Scientific Breakthroughs." At the opening main forum, two scientific masters took the stage one after another to share their views. Although their perspectives differed, they pointed to a common proposition: as AI continuously expands the boundaries of scientific research capabilities, humanity needs to rethink the methods and meaning of scientific discovery. This kind of intellectual convergence is precisely the value of the WLA Forum—scientists from different disciplines, different regions, and different generations examining the future of science under one roof. Beyond academia, this collision of top-level wisdom also reminds us that no matter what AI may change, curiosity about the unknown world, the pursuit of scientific truth, and the shared belief in open cooperation and benefiting humanity remain important forces driving science forward. Scene from the opening of the 2026 World Laureates Forum. Photo by Wen Hui Bao reporter Yuan Jing. When answers are no longer scarce, "choice" becomes key. In the long history of science, discovering new things often meant breaking through limits on resources and capabilities. But today, AI can rapidly generate massive numbers of research proposals, and scientists have suddenly been sent into a world with unprecedented abundance of "possibilities." But does having more choices make it easier to discover the unknown? Omar Yaghi, winner of the 2025 Nobel Prize in Chemistry, said in his speech that scientific research is moving from "scarcity" to "abundance," which brings a new question: when the possibilities available for exploration far exceed the scope humanity can study, how should scientists choose? The metal-organic framework materials Yaghi studies are the key to understanding this question. By carefully designing molecular structures, scientists can create new materials with different functions—for example, Yaghi's team focuses on special structural materials that can capture carbon dioxide and water from the air—but the potential combinations of molecular building blocks are astronomical. Powerful AI can easily generate tens of thousands of candidate materials, but which ones are worth further verification? In the past, scientists mainly relied on knowledge, experience, and intuition to choose candidate research targets. Yaghi goes further, proposing that a "science of choice" should be developed, making choice itself a scientific question that can be systematically studied and continuously improved. The significance of this concept goes beyond material screening: what questions to study, what experiments to conduct, and even how to promote applications in the future all cannot escape "choice." The core is to use AI to connect a broader range of disciplines, discover new laws and principles from them, and make scientific decisions evidence-based. Yaghi's laboratory is putting this "science of choice" into practice: after providing the necessary information and prompts to AI, the large model quickly produced 96 sets of crystallization conditions for potential new materials, which were then handed to robots and automated experimental platforms for verification. Soon, the results were fed back to AI and used to guide the next round of attempts... In this way, through several rounds of iteration, in less than a month, the crystallization performance of the material improved significantly, whereas traditional trial and error often takes years or even longer. This change is transforming the research experience of other participants. Nicholas Dirks, president and CEO of the New York Academy of Sciences, said that AI may be the most powerful engine humanity has ever created, and it is accelerating breakthroughs and discoveries in medicine, materials science, and basic science. Xie Weidi, a professor at Shanghai Jiao Tong University and winner of this year's WLA Prize Young Scientist Award, remarked that AI makes it possible to explore many topics that were once far out of reach. At present, he is trying to apply AI to research on rare diseases to further understand life mechanisms. For Yaghi, he believes that in the AI era, the "science of choice" is expected to become an important scientific method, and judging what is worth exploring will become a key capability in scientific research. Scene from the opening of the 2026 World Laureates Forum. Photo by Wen Hui Bao reporter Yuan Jing. When machines learn to prove, what is creation? The shock brought by AI is not limited to laboratories. The mathematics community, long known for rigorous logic and rational thinking, is generally facing a new confusion: AI is increasingly good at mathematics, but mathematicians still cannot understand the mechanisms behind large models. Laurent Lafforgue, winner of the 2002 Fields Medal, recalled that progress in machine translation once made many mathematicians feel similarly surprised. They had assumed that to make machines translate accurately, it would first be necessary, as in mathematical research, to establish a set of theories and rules; but using one set of theories to explain humanity's extraordinarily complex languages is almost impossible, so mathematicians believed machine translation could not succeed. Today, however, they not only have to retract that judgment, but large language models have left mathematicians dumbfounded. This inevitably prompts the question: if machines become increasingly good at solving problems, where should mathematicians' creativity be reflected? In Lafforgue's view, mathematical creation is far more than problem solving; it can have more dimensions: for example, creating new concepts, opening up new angles of observation, and even connecting originally scattered knowledge to form an entirely new understanding of the whole field of mathematics. A typical example is the creation of analytic geometry—it is less a matter of solving problems than a definition of an entirely new mathematical perspective. Michael Levitt, winner of the 2013 Nobel Prize in Chemistry, also said that when AI allows scientists to complete in an afternoon a concept that might once have taken a year to verify, what is truly scarce is the ability to ask high-quality questions. In fact, whether it is the posing of new scientific questions or the creation of new mathematical concepts and theories, scientists are all emphasizing that as AI makes solving easier, scientific research also needs to place greater emphasis on the value of the questions themselves and the ability to open up new directions. Laforgue's thinking goes even further: since AI can handle language so effectively and conquer mathematics, could there be mathematical structures hidden behind the complex texts that have not yet been fully recognized? At present, he is trying to find new mathematical theories to reveal these mysteries. This also shows, from one angle, that AI and mathematics still have room to improve each other. The 2026 Top Scientists Forum was lively with on-site exchanges. Photo by Yuan Jing, Wenhui Daily reporter. In an era of paradigm change, the foundation of science remains. The tools and methods of scientific research change with the times, but why science sets out and what it is ultimately for may not change because of the emergence of AI. The interpretation of the scientific spirit and the insistence on the foundation of science have always been a "hidden theme" of the Top Scientists Forum. In the context of artificial intelligence, scientists' curiosity about the unknown, their insistence on truth, and their sense of responsibility toward science are still unchanged in value. Yuan Junying, winner of this year's Top Scientists Award in Life Science or Medicine, recalled when accepting the award on stage that 44 years ago, it was a simple question—how do cells die—that led her to begin decades of exploration. Today, she still reminds young scientists not to blindly chase hot fields, but to always preserve curiosity, dare to blaze new paths, and pursue the "why" that comes from within. Reviewing the research journeys of this year's Top Scientists Award winners, Levitt said that whether it was wavelet theory or programmed cell death, neither was initially aimed at developing products, but rather arose from scientists' exploration of an interesting question. Only decades later did the application value of these basic research findings gradually emerge. "Curiosity comes first, and applied results will follow." Scientific progress cannot be separated from trust and openness. Dunning-Kruger pointed out that the new paradigm of scientific research means not only new tools and new discoveries, but also innovation in scientific cooperation. In the face of the opportunities and risks brought by AI, such cooperation is even more urgent. He particularly emphasized that, as important forces driving AI development, China and the United States bear special responsibility and should jointly discuss technical safety and governance to ensure that this remarkable technology can serve human well-being. "Cooperation and exchange must never be a luxury in the scientific community; they are the core way science operates," Dunning-Kruger said, noting that this is especially urgent in the field of artificial intelligence.