AI keeps cracking unsolved math problems, and mathematicians have mixed feelings

AI models are increasingly helping solve open math problems, including finding counterexamples and generating proofs.

Work highlighted by Epoch AI and OpenAI shows AI can deliver real results on unsolved research tasks, not just routine assistance.

Mathematicians are split between excitement over productivity gains and concern that AI could reshape how the field is practiced and understood.

The trend could accelerate mathematical discovery, but it also raises questions about verification, expertise, and academic culture.