Published August 21, 2026 · Written for Computer-science deans, curriculum committees, teaching centers and employers
Computer-science departments are being asked to teach two different AI literacies at once: deeper technical competence for their own students and practical AI capability for the rest of the university. In Primary Research Group's survey, 53.62% of computer-science faculty said their department was on par with peers in helping students use AI for coding. Another 20.29% believed their department was behind, while only 8.70% considered it ahead. The figures suggest that many departments are adapting, but relatively few feel they have established a clear lead.
The mandate extends beyond computer-science majors. Just over half—50.72%—agreed that their department should offer internal AI workshops for faculty and PhD candidates outside computer science; 17.39% strongly agreed and 33.33% agreed. Only 10.15% opposed the idea.
Universities have not approached that responsibility consistently. Some 31.88% of respondents said their institution had made a great deal or a lot of effort to promote AI-related courses for other disciplines, while 26.08% reported little or no effort. Asked who should be responsible, 37.68% pointed to the computer-science department and 24.64% to the disciplines needing the training. Nearly 29% gave no answer, a sign that ownership itself remains unsettled.
For academic leaders, the practical model may be shared responsibility. Computer-science faculty can supply technical foundations, model limitations and evaluation methods; disciplinary faculty can define authentic uses, domain standards and the consequences of error. Teaching centers and libraries can help scale common instruction on source evaluation, privacy and responsible use.
The curriculum question is no longer whether students should encounter generative AI. It is which capabilities belong in core computing courses, which should be embedded across disciplines and who will maintain the training as tools change.