UC Berkeley Computer Science Failure Rates Double as AI Tool Use Rises

UC Berkeley's intro computer science courses saw failure rates jump sharply in spring 2026: 35.3% in CS 10 and 10.6% in CS 61A, up from under 10% historically. Student GPAs fell to 2.3, below departmental targets of 2.8–3.3. Upper-division courses also struggled. Faculty point to two causes: students leaning too heavily on AI coding tools and gaps in basic math preparation entering the program.