AI-Driven Predictive Analytics for Personalized Learning and Early Academic Risk Detection

This study proposes an artificial intelligence (AI)-driven predictive analytics framework for personalized learning and early academic risk detection in digital education settings. The framework integrates machine learning models to estimate probabilistic risk, agentic AI to trigger timely and individualized interventions, and explainable AI to support transparent decision-making. In addition, responsible AI practices are embedded to address fairness, privacy, and human-in-the-loop oversight. By combining behavioral indicators (e.g., engagement regularity, time-on-task, and submission timeliness) with academic performance signals, the approach supports earlier and more actionable identification of at-risk learners than performance-only methods. The proposed design highlights how predictive insights can be translated into practical, ethically grounded support for instructors and learners.