Artificial intelligence is changing how universities teach and evaluate students, but policies and training have not kept pace. A recent study from South Africa’s Vaal University of Technology found lecturers are integrating AI tools faster than institutions can establish guidelines, creating gaps in accountability, ethics, and digital skills.
Policy lags behind AI adoption in classrooms
Research presented at the Second International Conference on Climate Resilient, Smart and Sustainable Futures in Zimbabwe this August surveyed VUT lecturers about their use of AI in teaching and grading. Many expressed uncertainty over when and how the technology should be used, describing institutional rules as inconsistent or absent.
“Lecturers reported confusion around accountability, ethical boundaries, and the use of AI-generated content,” the study noted. Without clear policies, instructors must decide whether AI belongs in assignments, exams, or student submissions—and how to detect misuse.
Related: ICT council chairperson removed from post
This situation mirrors a broader trend. Universities globally are struggling to define AI’s role in academia, especially as generative tools like chatbots produce essays, code, and research summaries with growing sophistication. In South Africa, some institutions have abandoned AI detection software due to reliability concerns, while others never adopted it.
Digital literacy gaps slow progress
The VUT study also revealed that limited digital skills prevent some lecturers from using AI effectively. While many view the technology as a way to improve efficiency, provide better feedback, and enhance course design, others lack the confidence to integrate it on their own.
Naledi Kaeane, the study’s lead researcher and a lecturer in VUT’s Department of Tourism and Integrated Communication, stated that AI is not replacing educators but reshaping teaching, assessment, and student engagement. She stressed that structured support and clear policies will determine whether the shift is managed responsibly.
Related: Rand makes debut on Luno Global
The findings suggest universities should focus on improving lecturers’ digital literacy rather than banning AI. Policies must define acceptable use, protect academic integrity, and redesign assessments to account for AI assistance. The aim is to ensure technological progress does not weaken trust in academic qualifications.
The issue extends beyond technology to governance. Universities need frameworks to distinguish between helpful AI use and misuse, including rules for disclosing AI-assisted work and assigning responsibility when it appears in submissions. Without these, the boundary between tool and shortcut remains unclear.
For now, the gap between adoption and policy remains significant. Lecturers are already using AI in classrooms, but without consistent rules, risks like plagiarism, privacy violations, and unreliable assessments could outweigh the benefits. The challenge is not just keeping up with the technology but deciding what kind of education system it should serve.
