Adaptive Learning in Online Education Using Artificial Intelligence: A Systematic Survey of Techniques, Challenges, and Future Directions
Abstract
Online learning platforms have significantly transformed modern education by providing flexible, scalable, and accessible learning opportunities. However, conventional e-learning systems continue to follow a one-size-fits-all model, delivering uniform content regardless of individual learner needs, pace, or prior knowledge. Adaptive learning has emerged as a promising solution by leveraging Artificial Intelligence (AI), Machine Learning (ML), and learning analytics to personalise learning experiences based on user behaviour, progress, and preferences. This paper presents a systematic survey of recent adaptive learning techniques in online education, reviewing rule-based systems, classification models, recommendation systems, deep learning, reinforcement learning, and Explainable AI (XAI). A comparative analysis examines their strengths, limitations, and suitability for personalised education. To address identified challenges, a four-layer conceptual framework - the AI Adaptive Learning Pipeline (AALP) - is proposed, integrating learner profiling, intelligent recommendation, explainability, and continuous feedback. Persona-driven design validation, wireframing, and usability evaluation outcomes are also discussed. Future research directions including federated learning, LLM-based tutoring, and multimodal analytics are presented.