Artificial Intelligence Solves Unsolvable Mathematical Problems

The resolution of previously unsolved math problems by artificial intelligence models has sparked debate among leading mathematicians regarding how future research will be shaped. OpenAI provided solutions to 719 problems last week, while the Simons Foundation announced a $15 million program to evaluate artificial intelligence results.

The recent breakthrough by artificial intelligence models in overcoming previously unsolvable problems in recent months has prompted high-level mathematicians to reflect on their future roles. Hugo Duminil-Copin, recipient of the 2022 Fields Medal, stated that research paths are changing. Artificial intelligence firms have been publishing solutions to unresolved mathematical problems since April to prove their model capabilities; indeed, OpenAI shared the solutions to 719 problems last week.

Stephane Mallat, recipient of the 2025 Gold Medal, argues that mathematics remains a central tool for clear thinking. While Francois Charles, head of the mathematics department at the Ecole Normale Superieure in Paris, adopts an optimistic approach regarding the scientific direction of these changes, Cedric Villani emphasized that the path to reaching a result is at least as important as the result itself.

On the other hand, the Simons Foundation announced a $15 million program in partnership with OpenAI and Anthropic to support mathematicians in evaluating artificial intelligence results and establishing standards.