UC Berkeley CS Courses Post 35.3% Failure Rate as AI Reliance and Math Gaps Deepen
Updated
Updated · VnExpress International · Jun 13
UC Berkeley CS Courses Post 35.3% Failure Rate as AI Reliance and Math Gaps Deepen
3 articles · Updated · VnExpress International · Jun 13
Summary
Spring 2026 failure rates jumped to 35.3% in COMPSCI 10 and 10.6% in COMPSCI 61A, far above the sub-10% levels seen in the prior two spring terms.
Average grades in both introductory courses fell to about 2.3 on a 4.0 scale from roughly 2.8 to 3.3, while the department’s guideline targets combined D and F rates below 7%.
Dan Garcia, who teaches both classes, said heavy use of ChatGPT, Claude and Gemini on assignments left many students unable to perform on proctored exams; nearly 30 CS 10 students were caught cheating on take-home work.
Faculty also pointed to weak math preparation, with instructors reporting struggles in basic linear algebra despite course expectations that students already know linear algebra, vector calculus and proofs.
More than 1,400 UC faculty have signed a petition to restore SAT or ACT requirements for STEM programs, citing a UC San Diego report that students entering with below-middle-school math skills have risen nearly 30-fold since 2020.
With student skills declining, is abandoning standardized tests a failed experiment for America's top universities?
AI was meant to make us smarter. Why is it being blamed for a historic academic failure at UC Berkeley?
UC Berkeley CS Failure Rates Triple to 35%: Math Deficiencies and AI Overreliance Expose Higher Ed Crisis
Overview
UC Berkeley's introductory computer science courses are facing a sharp rise in student struggles and failure rates in Spring 2026. Faculty observations reveal that many students lack essential mathematical foundations, such as linear algebra and proofs, which are assumed prerequisites for these courses. This gap between what students are expected to know and their actual skills is causing significant performance issues. Additionally, an overreliance on artificial intelligence tools is making it harder for students to develop the deep understanding needed for success. As a result, more students are entering advanced coursework unprepared, leading to a growing academic crisis.