Students’ mixed feelings about AI: evidence from four Australian universities

Updated Sep 07, 2026

Updated, 7 September 2026: We have consolidated two briefings on this paper, corrected the description of 50.0% as “more than half”, and made the sample and interpretation limits explicit.

What the study reports

Glenys Oberg, Yifei Liang, Margaret Bearman, Tim Fawns, Michael Henderson and Kelly E. Matthews examine students’ emotional experiences of AI in Feeling AI, published in Higher Education on 28 March 2026.

The research combines 8,021 survey responses with 79 participants in 20 focus groups across four Australian universities. Optimism and excitement could coexist with scepticism and worry. Focus-group accounts connected AI with questions about acceptable assistance, assessment and authorship.

The authors interpret these accounts through the institutional and moral context of AI use. They advocate discussion of its emotional dimensions alongside technical literacy. This is an argument informed by the findings, not evidence that a particular discussion programme improves trust or learning.

The participants were self-selected. The study captures a particular period in Australian higher education and does not estimate the views of all students, in Australia or the UK.

How a university could use this

Our suggestion is to ask separate questions about usefulness and confidence in the rules. A student’s answer to one need not answer the other. Invite examples of where course guidance was clear, where it was ambiguous and what clarification would help.

Keep experience reports separate from determinations of academic misconduct. A comment about anxiety is a reason to understand the student’s situation; it is not evidence that the student broke a rule. Likewise, favourable comments about a tool do not establish that it improved learning.

Before combining responses, record the course, task and guidance students were responding to. A local review can then identify which questions need a policy answer and which need a teaching discussion. These are our proposed review steps, not a tested method from this paper. See the student comment analysis governance checklist for planning an appropriate analysis workflow.

Reference

Oberg, G., Liang, Y., Bearman, M., Fawns, T., Henderson, M. and Matthews, K. E. (2026). Feeling AI: Circulating emotions, institutional climates, and moral boundaries in student use of AI. Higher Education. Read the paper. DOI: 10.1007/s10734-026-01658-6.