The Use and Usefulness of GenAI in Higher Education: Student Experience and Perspectives

Synthesis: Chung, Henderson, Slade, Liang, Pepperell, Corbin, and colleagues (2026) report a large-scale multi-institution survey of over 8000 students at four Australian universities on their GenAI use. Over 80% of students had used GenAI for study-related tasks, with nearly half using it regularly, primarily for editing, summarizing, and idea generation. Usefulness rather than blind Trust or rule compliance appeared to drive engagement: students adopted GenAI pragmatically, with low trust in its factual accuracy but confidence in managing it as a support tool. Most students learned about the tools informally, and equity gaps were evident, with lower engagement among women, non-binary, neurodivergent students, and those with health conditions.

Key Findings

• Over 80% of students at four Australian universities had used GenAI for study-related tasks, with nearly half using it regularly, mainly for editing, summarizing, and idea generation.
• Usefulness, rather than blind trust or rule compliance, appeared to drive engagement; students had low trust in factual accuracy but confidence in their ability to manage the tools.
• Most students learned about GenAI through informal, self-directed means rather than institutional resources.
• Equity gaps were evident: proportionately lower engagement among women, non-binary, neurodivergent students, and those with health conditions.
• While most students avoided prohibited use, a significant minority reported behavior that contravenes policy, pointing to a need for nuanced, situationally aware academic-integrity approaches.

What this means for practice

• Instructors. Build guidance around how students already work — editing, summarizing, and idea generation — rather than around rule compliance, since usefulness rather than Trust or policy drove engagement.
• Instructors. Make AI use discussable in class: students reported low trust in GenAI's factual accuracy but confidence in managing it as a support tool, so the teaching task is calibration rather than prohibition.
• Administrators. Target equity gaps directly, since engagement was proportionately lower among women, non-binary, neurodivergent students, and students with health conditions.
• Administrators. Fund institutional AI Literacy provision, because most students learned about GenAI informally through self-directed means rather than through university resources.
• Administrators. Rewrite Academic Integrity policy to distinguish prohibited use from the pragmatic, situationally negotiated use that a significant minority reported, instead of relying on blanket bans.

Limitations

• Respondents were N = 8021 students who completed the full survey, drawn from 10,132 volunteers (5% of roughly 192,000 invited students) distributed as 3859 from Monash University, 2582 from University of Queensland, 997 from University of Technology Sydney, and 583 from Deakin University.
• Recruitment was restricted to the researchers' respective universities, producing a convenience sample, and each institution applied its own distribution process.
• The design is a single cross-sectional online survey; frequency and purpose of use are self-reported rather than observed, and a further 826 students completed only part of the survey.
• Findings come only from four Australian universities, so the demographic and equity patterns cannot be assumed to transfer to other national systems or institution types.

Citation

Chung, J., Henderson, M., Slade, C., Liang, Y., Pepperell, N., Corbin, T., et al. (2026). The use and usefulness of GenAI in higher education: Student experience and perspectives. Computers and Education Open, 10, 100347.