Professor Richard Borcherds Stopped Grading Homework Due to AI

Professor Richard Borcherds from the University of California, Berkeley, has refused to check students' homework for correctness. He explained his reasoning in an interview with the California Post: the widespread use of AI has made such verification pointless.

Borcherds is well-known in the scientific community: in 1998, he received the Fields Medal, one of the most prestigious awards in mathematics, for proving the monstrous moonshine conjecture, which links the largest sporadic simple group, the Monster, with modular functions.

"Grading for correctness only creates a huge incentive to use AI for homework instead of doing it independently and punishes students for honesty," he said.

What has changed

According to the mathematician, using AI for assignments has become commonplace among students, and detecting such work is almost impossible: services for recognizing generated text often make mistakes and can accuse a student who did everything themselves of cheating. "There is no simple way to distinguish work done by a human from that generated by AI," he noted.

In a conversation with the university's student newspaper, the Daily Californian, Borcherds added that there is no point in paying graders to review homework. Now, points are awarded only for submitting assignments on time, and the correctness of answers does not affect the grade.

The professor does not have a ready solution for educators. He admitted that he tried to find an answer from AI itself, but unsuccessfully. "I have no idea what to suggest. I even asked ChatGPT this very question, and even ChatGPT replied that it doesn't know what will happen next," Borcherds said.

The problem is broader than one university

Educators worldwide regularly report on work clearly written with the help of ChatGPT and other chatbots. Some services rent out smart glasses that students use during exams.

Some countries and regions have begun to restrict the use of AI in education. In Norway, elementary school students are effectively banned from using AI, while high school students are allowed to work with it only under teacher supervision. In New York, restrictions were introduced in schools after parents and educators opposed humanoid AI robots being involved in children's education.

Borcherds' approach is the opposite of bans: instead of fighting AI, the professor simply stopped grading what students could delegate to a chatbot.