Design and Evaluation of a Personalized AI Tutoring System Using ChatGPT

Abstract

Personalized learning is challenging because students have diverse backgrounds, abilities, needs, and traditional teaching often lacks timely, tailored support. This study introduces an AI tutoring system powered by ChatGPT, designed to provide adaptive, context-aware educational assistance. The system uses conversational interfaces, domain-specific prompts, and learner modeling to adjust dynamically to each student's progress and misconceptions. An experimental study with high school and early university students showed that the system significantly improved learner engagement, comprehension, and efficiency. Users appreciated the personalized explanations and adaptive responses, which boosted their motivation and deeper understanding of the material. These findings demonstrate the potential of AI-powered tutoring systems to enhance personalized education by offering scalable, effective, and engaging learning experiences suitable for diverse learners and contexts. This approach could help overcome limitations of traditional education and support individualized learning .