Raising Ethical Awareness of GenAI Use Through Student Self-Assessment in the Transition to Higher Education

Synthesis: Pedlow and Maldon (2026) investigate how guided reflection embedded within Self-Assessment can support students to engage responsibly and confidently with generative AI (GenAI) in learning and Assessment, while maintaining academic integrity. Conducted at an Australian university (Edith Cowan) with commencing undergraduate and postgraduate students across nursing, health sciences, engineering and science (2021–2025), the study implemented pre- and post-semester self-assessments combining Likert-scale confidence items with an open-ended prompt inviting reflection on the ethical implications of GenAI use. Findings indicate that students recognized both the benefits and limitations of GenAI, demonstrating growing ethical awareness and critical evaluation of its role in academic work. Reflections revealed persistent tensions between efficiency and academic integrity, uncertainty about institutional expectations, and a shift from risk aversion and fear toward deliberate boundary-setting, disclosure, and critical vigilance. Students reported an increased sense of personal responsibility, describing the reflective process as confidence-building during their transition to higher education. The authors frame ethical GenAI use not as compliance but as a developmental capability, positioning guided reflection within Self-Assessment as a scalable, student-centered pedagogical strategy that fosters readiness, self-regulated learning, and learner agency.

Key Findings

• Guided reflection raised ethical awareness: students moved from pre-intervention risk aversion and fear (e.g. "I have never used GenAI as I'm too scared of breaching academic integrity") toward more defined strategies for ethical engagement around disclosure, accountability, and critical vigilance over outputs.
• Confidence in academic integrity increased most strongly across all domains (largest gains in referencing and paraphrasing), while confidence in using GenAI for learning and Assessment showed more modest gains — gains were largest among students with the lowest initial confidence.
• Undergraduates vs. postgraduates diverged: undergraduates showed the strongest gains in academic integrity skills but decreased confidence in using GenAI for learning/assessment (a realignment of understanding toward its limits and complexities), whereas postgraduates improved consistently from a lower baseline in assessment-related GenAI use.
• Tensions between efficiency and integrity persisted: reflections surfaced ongoing concerns about referencing, acknowledgment, fairness, reliability, hallucinations, and intellectual property — procedural and ethical complexity that a reflective intervention alone did not resolve.
• Reflection fostered learner agency and personal responsibility: students described the reflective process as confidence-building during their transition to university, linking ethical use to professional identity and resisting over-reliance on GenAI.
• Self-Assessment reframes ethical inquiry as pedagogy: embedding GenAI-focused reflection within Self-Assessment shifts the focus from compliance and detection toward critical, self-regulated engagement, offering a scalable model adaptable across disciplines and year levels.

Study Design & Method

The study used a pre-test/post-test quasi-experimental design (Rogers & Revesz, 2019) underpinned by a descriptive qualitative methodology (Braun & Clarke, 2013). Participants were commencing undergraduate (n=491) and postgraduate (n=206) students at Edith Cowan University recruited via convenience sampling across nursing, health sciences, engineering and science. The intervention was an Self-Assessment survey, co-designed with students and staff, originally developed in 2021 for academic and digital literacy and expanded with GenAI-specific questions in 2023. It addressed four domains: (1) confidence in academic integrity practices, (2) confidence using GenAI for Assessment preparation, (3) confidence using GenAI for learning, and (4) an open-ended ethical reflection ("Have you considered the ethical implications of using GenAI? If yes, how?"). Quantitative items used a four-point confidence scale and were analyzed descriptively with paired-samples t-tests on matched responses (n=25); open-ended responses were analyzed using Braun & Clarke's six-phase inductive thematic analysis with peer debriefing for credibility. Pre-surveys (n=697) were administered in Orientation/Weeks 1–2; post-surveys (n=67) in Weeks 12–13. Instrument reliability was excellent (Cronbach's α = .94 pre, .93 post). The authors note high pre-test but variable post-test response rates as a limitation, particularly for the matched subsample.

What this means for practice

• Instructors. Embed a short confidence-rating plus open-reflection Self-Assessment in orientation or the first weeks of a unit and repeat it at semester end, so ethical inquiry about GenAI becomes part of academic development rather than an externally imposed rule.
• Instructors. Teach the procedural conventions explicitly alongside ethical reflection — referencing, acknowledgment, and record-keeping — since confidence in academic integrity rose most in exactly those areas.
• Instructors. Read the responses to locate students starting from the lowest confidence, where gains were largest, and treat the reflective prompt as a low-cost way to surface deliberation about boundaries, disclosure, and accountability.
• Faculty developers. Frame academic integrity work as educative and student-centered rather than compliance- and detection-driven, and scaffold metacognitive awareness and self-regulated learning for GenAI-mediated environments. Note the paper's CC BY-ND (no derivatives) license before adapting its materials.

Limitations

• The matched pre/post analysis rested on only 25 paired responses drawn from 697 pre-surveys and 67 post-surveys; the authors note consistently high pre-test but variable post-test completion, a common attrition problem in voluntary survey research.
• Confidence was self-reported on a four-point scale and the open-ended prompt captured stated reasoning only, so no behavioral or outcome measure of actual academic integrity practice was collected.
• Data came from one Australian university (Edith Cowan) and a convenience sample of commencing nursing, health sciences, engineering, and science students, split unevenly between undergraduates (n=491) and postgraduates (n=206).
• The instrument was designed in 2021 and expanded with GenAI-specific items only in 2023, so cohorts across the 2021–2025 window did not answer identical questions.

Citation

Pedlow, M., & Maldon, J. (2026). Raising Ethical Awareness of GenAI Use Through Student Self-Assessment in the Transition to Higher Education. Journal of University Teaching and Learning Practice, 23(5).