UC Berkeley Professor Drops Homework Grading Due to AI Cheating Concerns

Mathematics Professor Shifts to Traditional Assessments

Professor Richard Borcherds of the University of California, Berkeley, has announced that he has stopped grading homework assignments for his mathematics students. According to reports, the decision stems from the increasing difficulty of distinguishing between human-written submissions and those generated by artificial intelligence tools.

The renowned mathematician indicated that the quality of AI-produced homework has advanced to a point where traditional grading methods are no longer effective. Consequently, student assessment in his courses will now rely entirely on traditional, in-person exams conducted without electronic devices.

Practical Implications and Context

• Shift in Evaluation: Assessments have moved away from take-home assignments toward supervised, offline testing.
• Detection Challenges: The university previously evaluated AI detection software and found it unreliable, leading the institution to allow individual professors to establish their own course policies regarding AI usage.
• Incentive Structures: Professor Borcherds noted that grading strictly for correctness creates an incentive for students to utilize AI tools to guarantee correct answers, inadvertently penalizing academic honesty.

How to Verify

To examine the official stance of educational institutions regarding artificial intelligence policies, interested parties can review course syllabi updates directly from the University of California, Berkeley, or consult academic policy guidelines published by the university administration.