Following the rapid rise of artificial intelligence, 24 leading mathematicians met at Harvard in September 2026 to address the crisis in training future researchers. The resulting report warns that the ability of AI to generate complex proofs threatens traditional academic assessment methods. Experts argue that a PhD should no longer rely solely on a written thesis, as the machine now handles much of the technical heavy lifting. Instead, they propose shifting focus toward the 'art of questioning,' collaborative interdisciplinary work, and rigorous in-person evaluations to ensure students truly grasp the material. The report emphasizes that while AI can accelerate understanding, it must not replace human insight. Researchers are urged to maintain strict ethical standards and oversight, noting that while the machine can solve equations, the ability to define significant research questions remains a uniquely human endeavor.