Academic Integrity in the Era of Generative AI: Students’ Perceptions, Ethical Boundaries, and Risk-Taking Behavior in Nigerian Universities

Abstract

Context: This study examined the intersection of academic integrity and Generative Artificial Intelligence (GenAI) adoption among students in Nigerian universities, addressing a critical gap in empirical, student-centered research from the Global South. Objective: To investigate students' knowledge, usage patterns, and perceptions of GenAI, as well as their awareness of academic integrity and the behavioral factors that shape ethical decision-making in AI use. Methods: A quantitative cross-sectional survey was conducted with 262 undergraduate and postgraduate students from nine Nigerian higher education institutions. The study was informed by relevant literature from major academic databases. Data were collected via a structured questionnaire and analyzed using descriptive and inferential statistics, with Prospect Theory applied as the theoretical framework. Results: Findings revealed high AI literacy, with 84.7% of participants already integrating AI tools into academic work. However, a significant knowledge–behavior gap emerged: while over 90% acknowledged the importance of academic honesty, only 36% believed AI use required disclosure. This ethical ambiguity was compounded by weak institutional guidance: 74.8% of students reported being unaware of their university AI policies. Inferential analysis indicated that students engage in risk–reward evaluations, where low perceived detection risks and academic pressures frequently outweigh potential sanctions. Conclusion: This study concludes that AI-related academic misconduct is often a rational behavioral choice driven by perceived institutional unpreparedness rather than ignorance. It calls for a transition toward adaptive academic integrity frameworks that prioritize ethical awareness and transparent academic policy communications to students.