Why eleven students described using GenAI around assessment

Updated Sep 07, 2026

Updated, 7 September 2026: We have made the small study’s scope explicit and removed claims that its findings establish support gaps or intervention benefits across institutions. This review uses the abstract and selected publisher passages.

What the source reports

Zara Hooley, Sitira Williams and Beverley Hancock-Smith interviewed 11 students at one UK post-1992 university. Recruitment covered different disciplines and levels, from first-year undergraduate to postgraduate study; data collection began in November 2023.

Participants described using text-generative AI to understand concepts, start assignments and refine wording. They valued immediate, private responses, sometimes more than accuracy or depth. Some accounts described seeking explanations that teaching had not provided in a usable form. The authors discuss these experiences through learner identity.

These are reported motivations and experiences. A small, varied sample does not make them representative, and perceived usefulness does not establish improved learning. We could not read the complete analysis and limitations. The revised account omits unverified quotations and does not infer that all AI use indicates unmet support needs.

Questions for course teams

Our suggested approach is to ask about the task and circumstances. What was the student trying to do? What help was available? What did the tool appear to provide, and what did the student check? A usage total cannot answer those questions.

Keep comments about convenience, confidence, explanation and assessment expectations distinct while allowing students to describe other reasons. Review any interpretation with the context of the activity. The governance checklist can help plan how sensitive accounts will be collected and used.

If feedback suggests a problem with access to teaching support, investigate that problem directly and assess any proposed response. More visible support does not automatically reduce AI use or improve attainment. These are our local review suggestions, not outcomes established by the study. Any wider claim needs evidence beyond these interviews.

Reference

Hooley, Z., Williams, S. and Hancock-Smith, B. (2026). Impact of text-generative artificial intelligence tools on students’ approach to assessment: a case study of a UK institution. Studies in Higher Education. Published online 17 April 2026. Publisher article.