Has AI made any progress as a mathematical theory-builder,,,,? (mathoverflow.net)

In a thought-provoking exploration, Timothy Gowers highlights the dichotomy in mathematics between problem-solvers and theory-builders, raising questions about the extent to which AI has progressed in both domains. While AI has demonstrated notable advancements in solving complex mathematical problems, the capacity of these systems to contribute to mathematical theory—such as introducing new concepts or frameworks—remains uncertain. Gowers calls for a deeper examination of whether AI can indeed make original contributions that reshape mathematical thinking or suggest novel research directions. This inquiry is significant for the AI/ML community as it addresses the limitations and potential of current AI technologies in advancing mathematical understanding. The lack of clear examples where AI has substantially influenced theoretical frameworks suggests a gap in our comprehension of AI's capabilities beyond problem-solving. By encouraging the documentation of both successful and unsuccessful attempts at AI-driven theory-building, Gowers invites the community to reflect on the originality of AI-generated concepts and the implications of these contributions for the future of mathematical research.