AI use leads UC Berkeley math professor to see 'no point' in grading homework

Your homework assignments may no longer be graded for correctness, shifting the focus of assessment in some math courses.

Richard Borcherds, a UC Berkeley mathematics professor and Fields medalist, has stopped grading homework for correctness due to widespread AI use among students. He observed that many students were achieving top scores by using AI tools, making traditional grading of homework assignments seem pointless. Borcherds revealed this decision in an interview following news of AI agents solving complex mathematical problems like the Navier-Stokes existence and smoothness problem. He expressed concern about the future of mathematics as a career, noting that even ChatGPT could not provide a clear answer when asked about career prospects in mathematics. While the university encourages faculty to include AI use statements in syllabi, policies vary by course, and current AI detection tools are deemed inadequate for campus-wide standardization. Borcherds currently grades homework based on completion, acknowledging it's an imperfect system with no immediate alternative.

The increasing capability of AI in mathematics challenges established academic assessment methods, pushing educators to reconsider the value of traditional assignments like homework. This situation signals a potential shift in pedagogical approaches, where the emphasis may move from evaluating rote problem-solving to assessing conceptual understanding or AI utilization skills. The difficulty in detecting AI-generated work also highlights a broader technological arms race in academia, where universities struggle to adapt policies and tools to the rapid advancements in AI.

Borcherds's dilemma reflects a growing unease within academic fields about AI's impact on the future of human expertise and career paths. The fact that a renowned mathematician is questioning the viability of his field as a career, and even seeking advice from AI itself, underscores the profound uncertainty AI introduces. This uncertainty is likely to ripple through other disciplines, prompting a wider reevaluation of what skills will be valuable and how education should prepare students for a future where AI is a ubiquitous tool or collaborator.