Mathematics Meets A.I.: Insights from Seattle's Joint Mathematics Meetings

Bryna Kra, president of the American Mathematical Society, aimed to challenge attendees with the chosen theme, urging them to contemplate the implications of artificial intelligence in various aspects of mathematics, including teaching and research. “It’s time to start thinking about how A.I. impacts your teaching, your students, your research,” she emphasized in an interview. She indicated that mathematicians should consider the implications of having A.I. tools as co-authors in their work.A highlight of the event was a keynote address by Yann LeCun, the chief A.I. scientist at Meta. During his lecture titled “Mathematical Obstacles on the Way to Human-Level A.I.,” he humorously critiqued the current state of machine learning, suggesting, “the current state of machine learning is that it sucks.” He argued for the development of a “large-scale world model,” which could enhance A.I.'s reasoning capabilities by creating mental models that foresee the consequences of actions. Nevertheless, he acknowledged the challenges that remain, particularly with certain mathematically intractable problems that continue to hinder progress in achieving human-level A.I.This year’s J.M.M. thus served as both a gathering and a platform for critical discussions on the role of mathematics amid the rise of artificial intelligence, encouraging the community to prepare for a future where such technologies become increasingly integrated into their work. As the mathematics community looks ahead, the convergence of these two domains holds the promise of reshaping existing paradigms, but it also poses significant questions that experts will need to tackle in the years to come.