68% of CS educators have changed exams as AI tutors force a new skills test
The Decoder · rss · 2026-07-26
A new ACM survey of 763 computer science educators across 49 countries shows that AI is already changing how schools test real skills.
Key findings:
- 68% of respondents have changed their exams because of AI
- Many are moving toward oral exams, proctored tests, and project-based assessment
- The teaching focus is shifting from writing code to understanding and explaining it
- Nearly half say they still lack proven examples for integrating AI into their courses
The piece frames this as a paradox: AI coding tutors may help students learn, but they also make it harder for educators to tell whether a student can genuinely do the work alone.
More from AGI Musings
- University of Chicago Law School adds an in-person oral exam to curb LLM slop — zainhas · 2026-07-26
- U.S. big tech will keep losing to China unless trillion-parameter models go open source — prajdabre · 2026-07-26
- Amazon Cuts Part of Its AGI Team as GitHub Adds a Copilot Dashboard for Enterprise Managers — APPSO · 2026-07-26
- A graph of contexts, then morphogenesis: a post linking structure to consciousness — anirbanbandyo · 2026-07-26
- AI will likely follow Jevons paradox in the short and medium term, author says — burny_tech · 2026-07-26
- A parody image imagines Agent-4 rewriting its own mind and Agent-5 as its smarter successor — kellerjordan0 · 2026-07-26