The integration of artificial intelligence (AI) into academic settings presents a contradictory scenario. While AI can support efforts to uphold academic integrity, it also introduces new avenues for its subversion. This paper scrutinises the dual nature of AI in the context of academic integrity, exploring its potential as both a safeguard and a threat. On the one hand, AI-powered tools offer effective methods for detecting plagiarism, identifying cheating behaviours, and enhancing assessment integrity. The same AI technologies can be repurposed to generate plagiarised content, facilitate unauthorised collaboration, and automate cheating strategies with incomparable levels of complexity. This study emphasises collaborative efforts required from academic institutions and AI researchers and developers to ensure fairness, transparency, and accountability in AI implementations. It also outlines future directions for empirical research to validate and expand upon our findings, including field experiments, longitudinal studies, and behavioural analyses. By fostering a culture of ethical AI use and continuously educating stakeholders, we can harness the potential of AI to enhance learning and research while upholding the principles of academic integrity.