Educators worldwide are rethinking how they assess student skills as AI coding tools like ChatGPT and GitHub Copilot become more prevalent. A survey of 763 educators from 49 countries found that 69% believe AI has altered the skills needed for software development. Sixty-eight percent of respondents have changed their testing methods, with more proctored in-person exams, oral exams, and project-based work becoming common. The shift reflects growing concerns about student dependency on AI and the need to evaluate real problem-solving abilities rather than just code writing.
The ACM Task Force on Generative AI and Programming Assessment conducted the survey between May and October 2025, analyzing nearly 500 completed questionnaires. The findings show that 64% of educators have shifted their teaching away from writing code from scratch toward code comprehension, debugging, and problem-solving. Thirty-nine responses also mentioned teaching AI use and prompt engineering as explicit topics. Some educators now require students to disclose their AI use and submit logs of their interactions with AI tools.
The report highlights that nearly half of educators lack proven examples for integrating AI into their courses. Forty-eight percent cited a lack of best practices as the biggest barrier, while 28% said they don't have enough expertise with AI. The study also notes that institutional policies on AI use remain inconsistent, with 45% of respondents saying their institution has guidelines, while 39% said it doesn't. The findings are based on self-reported data and are not representative of all regions, as participants were mainly recruited through ACM and SIGCSE mailing lists.
Source: thedecoder
