Use of generative AI in database courses: a role-based review

Introduction Generative artificial intelligence (genAI), particularly large language models (LLMs), has attracted increasing attention for its potential to support teaching and learning in higher education. However, how genAI is pedagogically integrated into specific disciplinary contexts, such as database education, remains insufficiently understood. Method This review examines the use of genAI in database courses through a role-based analytical framework, distinguishing the instructional function of AI as supplementary assistant, direct mediator, and new subject. Following PRISMA guidelines, peer-reviewed journal articles indexed in Web of Science and Scopus were screened. From an initial pool of 406 studies, nine empirical studies were identified. Findings GenAI is most commonly used as a supplementary assistant in database courses, where it supports tasks such as querying, data generation, solution checking, and error correction. The studies reported positive effects on academic performance, instructional efficiency, and time savings, with relatively few ethical or pedagogical concerns. The use of AI as a direct mediator demonstrated clear benefits but also raised concerns related to the accuracy and reliability of AI-generated evaluations. Finally, one exploratory study positioning AI as a new subject emerged as the most transformative yet risky role. While this approach offers potential for substantial instructional innovation, it may also lead to challenges related to accuracy and equity. Discussion This review underscores the potential and importance of role-based integration of genAI in database education and identifies key directions for future research. Research on AI usage in database instruction is still at an early stage, and further empirical studies are required.