Communicating academic integrity policy in the age of generative AI

Abstract

This study aims to examine how students and teaching staff perceive and enact academic-integrity policies in Romanian universities under the normative uncertainty created by generative artificial intelligence (AI). Parallel online questionnaires were completed by 786 students and 472 teaching-staff members at four public universities; closed-item analyses were complemented by thematic content analysis of open responses. Teaching staff evaluated integrity infrastructure more positively, while students perceived more examination and AI-assisted cheating. Student reporting was rare; teaching staff reporting intentions followed hierarchical internal channels. Unacknowledged AI use was judged more contextually by students and more categorically by staff. Cross-sectional, non-probability samples, self-selection, limited student coverage, non-identical role-adapted measures and the absence of respondent-level institution identifiers restrict generalisation and preclude causal inference. The results provide universities with a basis for discipline-specific AI guidance, disclosure rules, proportionate sanctions and credible reporting routes that distinguish learning support from replacement of student reasoning or authorship. Differences between students and teaching staff regarding cheating, AI-powered essays, and the practices of peer reporting can seriously damage the overall credibility of academic standards and even undermine the value of university degrees altogether. Peer reporting rates remain extremely low, suggesting tolerance to cheating behaviour or reluctance in peer reporting. Having agreed-upon AI usage standards would increase both fairness and accountability. By linking policy translation, social norms and responsive regulation, the study offers comparative evidence from students and teaching staff from an under-researched Central and Eastern European setting.