AI-Driven Personalized Learning in Educational Systems: A Framework for Adaptive Learning and Decision Support

Abstract

Artificial intelligence (AI) has expanded the capacity of educational systems to support learner modelling, performance prediction, adaptive resource recommendation, automated feedback, and conversational tutoring. However, existing AI-based personalized learning systems commonly employ these capabilities as separate technical components, with limited coordination, pedagogical grounding, explainability, educator control, and governance. This study proposes the Pedagogically Constrained Explainable Orchestration (PCEO) framework to address these limitations. A structured literature review guided by PRISMA 2020 reporting principles and critical thematic synthesis was used to identify recurring technical, educational, ethical, and human-oversight requirements in AI-driven personalized learning. These findings were translated into a requirement-driven con-ceptual architecture. The proposed framework coordinates specialized agents for learner modelling, engagement and risk analysis, recommendation, policy optimisation, generative support, and governance through a shared learner-state representation and an explicit orchestration mechanism. Candidate educational actions are evaluated using pedagogical suitability, expected learning benefit, engagement, explainability, cognitive load, risk, evidence quality, and governance constraints. The framework also incorporates decision-level explanation records and educator review, modification, rejection, and escalation mechanisms. The principal contribution is not the intro-duction of new standalone AI models, but the specification of how established AI capabilities can be combined within a transparent, pedagogically informed, and human-governed educational decision process.together with a staged implementation and evaluation roadmap The PCEO framework remains a conceptual and technically specified proposal; no classroom implementation or empirical effectiveness evaluation was conducted. Future work should develop and assess a working prototype through expert review, usability studies, and controlled evaluations involving educators and learners.