Abstract
This study examines how awareness of generative AI academic integrity policies shapes student anxiety, ethical reasoning, and writing practices in higher education. Drawing on a reflexive thematic analysis of student discourse from YouTube and Reddit, the study explores how detection-oriented governance frameworks are experienced and negotiated outside formal institutional contexts. Findings indicate that students do not primarily describe intentional rule-breaking; instead, they articulate experiences of surveillance, uncertainty, and adaptation within assessment systems perceived as inconsistent and adversarial. Three interrelated themes are identified: surveillance anxiety and defensive authorship; normative dissonance and moral negotiation of AI use; and cognitive offloading, systemic strain, and declining trust in assessment integrity. Across these themes, AI policy awareness is shown to shape not only behavioural responses but also students’ affective and epistemic relationships with writing, authorship, and institutional authority. The study argues that detection-first approaches risk intensifying anxiety and undermining trust without resolving underlying tensions in assessment design. It concludes that generative AI exposes structural fragilities in higher education, suggesting the need to reorient academic integrity governance toward design-led, transparent, and equity-sensitive approaches that prioritise learning, trust, and student experience.
Keywords
Academic integrity, generative AI, student wellbeing, surveillance anxiety, cognitive offloading.
