Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes
4 hours ago
- #Math Underpreparedness
- #AI Cheating
- #University Policy
- Failing grades in UC Berkeley CS classes like CS 10 (35.3% F) and CS 61A (10.6% F) surged in spring 2026, far exceeding department guidelines of 7% D/F for lower-division courses.
- Instructors cite AI-driven academic dishonesty as a key factor, with nearly 30 students caught cheating in CS 10, and students over-relying on LLMs for assignments but struggling on exams.
- Mathematical unpreparedness among students is a concern, highlighted by issues in courses like EECS 127 (16.8% F rate) and reports of AI-permissive math prerequisites weakening skills.
- Faculty signed a petition to reinstate SAT/ACT for UC STEM admissions to address math preparation issues.
- Understaffing forced cuts like removing final projects in EECS 127, and fewer TAs due to budget constraints, reducing academic support.
- Low student engagement, with sparse office hours attendance despite encouragement, reflecting decreased interaction.
- Professors plan changes: advertising past failures, adding remedial support, and emphasizing critical thinking to prepare students for a competitive, AI-driven world.