EduAI Personalized Learning System: A Large Language Model-Based Adaptive Learning System to Enhance the Effectiveness of Digital Education

Abstract

Digital education has expanded access to learning resources, but conventional digital learning environments frequently provide standardized content that does not adequately accommodate differences in learners' knowledge, learning pace, interests, and proficiency levels. Recent advances in large language models (LLMs) provide opportunities to develop adaptive educational systems capable of dynamically generating learning materials, explanations, feedback, and learning pathways. This study aims to develop the conceptual architecture of EduAI Personalized Learning System, an LLM-based adaptive learning system designed to support individualized digital learning. The study employed a design-oriented developmental approach consisting of needs identification, literature-informed system design, adaptive learning architecture development, and functional analysis. The proposed system integrates learner profiling, learning-path recommendation, LLM-based content generation, formative assessment, personalized feedback, and continuous learner-model updating. The analysis indicates that LLMs can provide a flexible interaction layer for adaptive learning while analytics and learner models provide the personalization mechanism. However, issues concerning hallucination, bias, privacy, pedagogical alignment, and excessive dependence on automated tutoring require explicit governance mechanisms. The proposed EduAI architecture therefore combines generative capabilities with learner analytics, knowledge representation, assessment, and human oversight. The study contributes an integrated conceptual framework for implementing LLM-powered personalized learning in digital education and provides a foundation for subsequent empirical evaluation.