Consistent GenAI governance is associated with student trust and disclosure

Updated Sep 10, 2026

Updated, 7 September 2026: We have separated the reported associations from claims that a particular policy or feedback process causes safer disclosure.

What the study reports

Wanqing Xia and Wei Wei examine how the implementation of GenAI governance relates to students’ trust and transparency behaviours. Their study follows 739 students across 25 courses over four waves and uses multilevel regression.

The authors report that implementation fidelity—the consistency with which governance is put into practice—was a more consistent predictor of perceived procedural justice than policy clarity. Procedural justice predicted trust, and trust was associated with disclosure and concealment behaviours. Direct justice-to-behaviour paths were not statistically significant once trust was included.

The authors also compare students with and without direct contact with governance processes. The interaction had a direction consistent with their argument but did not meet conventional statistical significance. Implementation fidelity remained relevant in both groups. The evidence should not be presented as a confirmed difference between those groups.

Four waves add a time dimension, but do not by themselves establish causality. These associations do not show that clearer policy alone, more feedback collection or a particular analysis tool increases honest disclosure.

Questions for a local review

Our suggested starting point is to compare the written policy with students’ experience of its application. Can students explain what is permitted, how a decision is reached and where to seek clarification? Do similar situations receive consistent explanations?

If you ask about trust or disclosure, explain how responses will be used and who can see them. Keep students’ accounts separate from conclusions about actual misconduct. A comment about uncertainty may identify a question for the course team without establishing a hidden behaviour or its cause.

The comment-analysis governance checklist can support planning for a local review. These are our suggestions, not a tested intervention from this paper. Any proposed policy change needs its own assessment, including unintended effects and the experiences of students who do not respond.

Reference

Xia, W. and Wei, W. (2026). Beyond the syllabus: How the fidelity of GenAI governance implementation shapes student trust and transparency behaviours. Innovations in Education and Teaching International. Published online 21 April 2026. Publisher record.