Richard Borcherds, a UC Berkeley professor and Fields medalist, has stopped grading student homework because AI use makes grading for correctness impossible. After students consistently received top scores using AI, Borcherds felt there was "no point" in grading. He now only grades on completion, admitting the system is "not very good" but has no alternative.
Borcherds worries the math department is no longer a safe career from the AI boom, stating he does not know whether to encourage students to enter the field. He even asked Chat GPT about the future, which gave no clear answer. He compared the situation to physical paper journals that have been "utterly pointless" for two decades despite still publishing.
The Academic Senate recommends faculty include AI use statements in syllabi, but policies vary by course. Turnitin and Rumi detection tools have proven inadequate for standardized campus use. Faculty can use a growing database of sample AI statements from the AI Hub to guide their syllabi.
Source: Hacker News · Summarized by HeadlinesBriefing