AI-Driven Adaptive Learning Systems in Higher Education: A Systematic Review
Thanet Yuensook, Thada Jantakoon, Potsirin Limpinan
Abstract
This systematic review investigates the implementation and impact of AI-driven adaptive learning systems in higher education, based on an analysis of 15 empirical studies published between 2020 and 2024. Following the PRISMA 2020 guidelines, the review addresses four core questions: (1) research trends and geographic distribution; (2) types of AI technologies and system characteristics; (3) implementation strategies in educational contexts; and (4) effectiveness and challenges encountered. The findings indicate a substantial increase in publications after 2022, with 73% of the studies published in 2023–2024. Geographically, research contributions span 15 countries, with the United States, China, and Europe as leading contributors. The predominant AI technologies identified include machine learning (40%), natural language processing (33%), and hybrid systems (27%), supporting real-time personalization and adaptive feedback mechanisms. Implementation strategies were observed primarily in STEM fields, language learning, and hybrid learning environments, with applications ranging from intelligent tutoring systems to LMS-integrated AI assistants. Effectiveness outcomes reported academic performance gains of 15–25% and improved learner engagement by up to 40%. However, challenges persist, including insufficient technical infrastructure, faculty readiness, ethical concerns (e.g., data privacy, algorithmic bias), and the underrepresentation of non-STEM disciplines. This review highlights critical considerations for successful integration of AI-enhanced adaptive systems and provides strategic guidance for institutions aiming to enhance personalization, equity, and scalability in higher education.