AI Lifted Homework Scores, So a Berkeley Professor Now Grades on Submission, Not Correctness
The Daily Californian reports that UC Berkeley mathematics professor Richard Borcherds is awarding homework credit for turning the assignment in rather than for getting the answers right. Whether generative AI is allowed in a given class is up to the course and the instructor.
As more students use AI to earn high homework scores, a UC Berkeley mathematics professor has changed how he grades. According to the Daily Californian, professor Richard Borcherds now awards homework points based on whether the assignment was submitted, not on whether the answers are correct. The standard has shifted from “did you get it right” to “did you hand it in.”
How homework is scored now
Borcherds concluded that grading for correctness had lost its meaning as more students used AI to earn high homework scores. For students, what changed is the basis for receiving credit. The measure is no longer whether the submitted work and answers are correct, but whether the homework was turned in at all. Homework remains an assignment you have to submit in order to get credit.
Who decides whether AI is allowed
UC Berkeley leaves it to individual courses and instructors to decide whether generative AI is permitted in class. Students have to check the rules for each course they are taking.
The university’s center for teaching and learning advises instructors to state their AI policies clearly in the syllabus. The point is to make it possible for students to know whether they may use AI in a given class. The guidance includes examples both of cases where AI use is allowed and cases where it is prohibited, and actual policies do differ from class to class.
The questions this leaves for Korean students and educators
The questions are what to assess in homework produced with AI, and how to communicate that standard to students. Borcherds made submission the basis for homework credit, and UC Berkeley advises that course-level rules be spelled out in the syllabus.
What is worth watching next is how well submission-based grading reflects student learning. Only by looking at how students respond and how learning outcomes change after the shift in grading can the effect of this choice be judged.