AI-Driven Personalized Education: A Literature Review of Learning Challenges and Theoretical Foundations

Abstract

Brandon F. Atonte

This literature review examines the theoretical foundations and empirical evidence supporting AI-driven personalized education systems for children aged 3-17. The review analyzes current educational challenges revealed by international assessments, examines established learning theories that inform AI educational design, and synthesizes research on pedagogical agents and multimedia learning. The review identifies significant academic performance declines in mathematics, reading, and science across OECD countries, alongside complex patterns in cognitive development research. Theoretical frameworks, including Cognitive Load Theory, multimedia learning principles, and Self- Determination Theory, provide evidence-based foundations for AI educational interventions. Meta-analytic evidence suggests personalized learning systems can achieve effect sizes of 0.15-0.30 SD, with stronger impacts in mathematics than language arts.

PDF