AI Transforms Education: UC Berkeley Professor Stops Grading Homework
Renowned mathematician Richard Borcherds opts for in-person exams as AI makes take-home assignments impossible to verify.
A Return to Traditional Methods
The rapid integration of artificial intelligence into academia has forced university professors to face difficult choices. Richard Borcherds, a mathematics professor at the University of California, Berkeley, has officially announced that he will stop grading student homework. According to Borcherds, modern technology has reached a point where it is virtually impossible to distinguish between work completed by a student and that generated by AI.
Borcherds, recognized for his contributions to mathematics and his current research in quantum field theory, believes that grading homework has lost its purpose. Consequently, his students will now demonstrate their knowledge exclusively through in-person, proctored exams, where the use of any electronic device will be strictly prohibited.
Why Homework Has Become Obsolete
The professor argues that the core issue lies in "accuracy-focused assessment." When a student is required only to provide the correct answer, AI becomes a highly effective, yet destructive, tool. Borcherds notes that average scores during online exams have become suspiciously high, which directly points to the widespread use of AI.
"Grading homework only encourages students to use AI instead of working independently. This approach effectively punishes honest students, and spending time and resources checking AI-generated answers is absurd," the professor states.
The University's Stance and the Future
The Berkeley administration previously tested AI detectors, but the tools proved unreliable. As a result, the university has decided to grant professors the autonomy to determine their own policies regarding the use of artificial intelligence.
Richard Borcherds' decision may set a precedent for other academic circles. While some lecturers are attempting to adapt AI into the teaching process, others, including Borcherds, believe that traditional, controlled environments remain the only way to maintain the quality of education. This development is particularly relevant for students and academic staff globally, as similar challenges are already emerging in universities everywhere. The question of how to evaluate students to ensure their knowledge is genuinely tested—rather than just the efficiency of algorithms—will be one of the primary challenges for the education system in the coming years.