Impact of AI-Powered Personalized Learning on Students’ Individual Learning Progress

Abstract

Artificial intelligence-powered personalized learning is increasingly used to adapt educational content, pace, difficulty, and feedback to the changing needs of individual learners. Despite this expansion, much of the existing evidence evaluates overall class performance rather than measuring how individual students progress through distinct learning trajectories over time.
This study proposes a learner-centric framework for examining personalized learning through mastery growth, learning velocity, retention stability, and engagement-sensitive adaptation. The research focuses on the extent to which AI-generated recommendations can improve learning progress while reducing performance stagnation among students with heterogeneous prior knowledge. A longitudinal experimental design is conceptualized in which student interaction data, assessment histories, response latency, hint usage, and topic-level mastery are incorporated into an adaptive learning model. The proposed framework combines knowledge tracing, learner-state representation, and policy-based content recommendation to estimate when, what, and how difficult the next learning activity should be. The central contribution is the use of individualized progress trajectories rather than aggregate accuracy alone, allowing learning gains to be interpreted in relation to each learner’s starting level and adaptation history.