At the International Congress of Mathematicians 2026, mathematicians from around the world gathered to explore the growing influence of artificial intelligence on their field. AI systems are now capable of generating mathematical proofs, identifying connections between seemingly unrelated areas of math, and even solving problems that had long stumped human researchers. The speed of these developments has sparked discussions about the true capabilities of AI and what might be gained or lost as mathematics becomes more reliant on machine assistance. The event was featured on The Joy of Why, a podcast hosted by Janna Levin and Steven Strogatz that delves into major unresolved questions in mathematics and science. The panel included Akshay Venkatesh from the Institute for Advanced Study, Ravi Vakil from Stanford University and president of the American Mathematical Society, and Alex Kontorovich from Rutgers University. They discussed recent breakthroughs, such as AI models scoring highly in the International Mathematical Olympiads and solving longstanding problems like the Erdős unit distance conjecture. Venkatesh pointed out that while AI has made impressive strides, it does not equate to superhuman intelligence. He argued that mathematics is more than a competition—it is a pursuit of understanding and storytelling. Vakil agreed, emphasizing that progress in mathematics is not just about solving problems, but about the narratives and human insights that drive the field. He also warned that viewing mathematics solely as a competition could be detrimental, particularly for younger mathematicians. Kontorovich highlighted AI’s potential to uncover new mathematical ideas, such as the counterexample to the Erdős unit distance conjecture, which spurred further research. However, he also raised concerns about the risks of AI generating incorrect proofs, which could overwhelm mathematicians with the task of verifying them. Some researchers are even being offered payment to review AI-generated proofs, though many are hesitant to take on the responsibility. The panel also considered the impact of AI on mathematics education. Vakil expressed concern that the availability of AI tools might diminish the importance of traditional problem-solving and critical thinking, leading to a generation of students who rely on AI for answers rather than developing their own understanding. He compared this to past debates about the use of calculators in education, noting that while tools change how people learn, the core goals of education remain unchanged. Kontorovich used the analogy of a forklift to explain how AI could be used in education: as a tool to complete tasks, as a means of learning how to use the tool, or even as a way to engage in physical exercise. He argued that the goal of mathematical education should be to develop critical thinking and the ability to endure frustration, not just to find answers quickly. The discussion also touched on the philosophical question of whether AI can truly understand mathematics. Kontorovich suggested that while AI may not be conscious, its role in helping mathematicians explore new areas is undeniable. However, he acknowledged concerns about the long-term impact of AI on the field, particularly if it surpasses human capabilities in all aspects of mathematical research. The panelists concluded that while AI presents challenges, it also offers opportunities for new discoveries and insights. They stressed the importance of maintaining a human-centered approach to mathematics, valuing storytelling, human understanding, and the pursuit of knowledge for its own sake.