Survey: 68% of computer science educators changed exams because of AI

An ACM survey of 763 computer science educators across 49 countries found that 68% have already changed how they test students because of AI tools, shifting toward oral exams, proctored tests, and project-based work. Nearly half of respondents said they lack proven examples for integrating AI into their courses.

Why it matters: This quantifies a shift that had mostly been anecdotal -- CS education is moving from testing code-writing toward testing code understanding, a direct downstream effect of AI coding assistants becoming standard tools. It's a leading indicator for how other skill-based fields may need to redesign assessment as AI tools become normalized.