The Effect of AI-Driven Personalized Learning with a Recommendation System on Students' Conceptual Understanding and Analytical Skills: The Moderating Role of Self-Regulated Learning

This study investigates the effect of AI-driven personalized learning supported by a recommendation system on students' conceptual understanding (C2) and analytical skills (C4), with self-regulated learning (SRL) as a moderating variable. Although AI-based personalized learning has attracted growing attention in education, empirical evidence regarding its effectiveness in controlled learning environments that consider differences in students' SRL levels remains limited. This study addresses this gap by implementing a structured AI-supported personalized learning approach through a recommendation system. A pretest–posttest control group experimental design was conducted with 60 senior high school students classified into high- and low-SRL groups using z-scores. The experimental group participated in AI-supported personalized learning through a learning management system (LMS) integrated with a recommendation system, whereas the control group received personalized learning without AI support. Data were analyzed using a 2 × 2 factorial MANOVA after meeting the assumptions of normality, homogeneity, and initial equivalence. The results showed that AI-driven personalized learning significantly improves students' C2 and C4. Although SRL does not have a significant direct effect on either outcome, it demonstrated a significant interaction effect on C4. These findings suggest that AI-based recommendation systems serve as an effective scaffolding mechanism for learners with different SRL levels. This study provides empirical evidence that integrating AI into controlled personalized learning environments enhances students' C2 and C4 in mathematics learning.