Intelligent Tutoring Systems for Personalized Learning Pathways in Digital Pedagogy

Abstract

The integration of Intelligent Tutoring Systems (ITS) in digital pedagogy has transformed educational personalization through artificial intelligence-driven adaptive learning. This study examines ITS implementation, effectiveness, and pedagogical implications in creating personalized learning pathways. Through comprehensive analysis of literature and a prototype implementation, we explore how AI-driven systems optimize educational outcomes by adapting content, pacing, and instructional strategies to learner characteristics. The research investigates key ITS components including knowledge representation, learner modeling, pedagogical strategies, and interface design. We present a functional prototype implementing Bayesian Knowledge Tracing (BKT) for algebra tutoring, which was evaluated in a controlled experimental study with 127 undergraduate students at Namangan State Pedagogical Institute. Our prototype evaluation shows an effect size of Cohen's d = 0.91, with the experimental group achieving a learning gain of 18.3 points compared to 14.1 points for the control group (p < 0.001). Furthermore, previous studies in the literature suggest that ITS can lead to an approximately 23% improvement in test scores and a 31% reduction in acquisition time in large-scale deployments. However, challenges remain related to design complexity, infrastructure requirements, teacher training, and ethical considerations. This article proposes a framework for effective implementation of ITS and discusses future research directions.