Artificial intelligence in university mathematics education: Personalized tutoring support and learning analytics as factors of academic achievement, independent thinking and risk reduction

Roman Yavich

African Educational Research Journal
Published: March 23 2026
Volume 14, Issue 1
Pages 228-239
DOI: https://doi.org/10.5281/zenodo.19183378

Abstract

In recent years, artificial intelligence has been increasingly integrated into higher education, including the teaching of mathematics. However, its impact on students’ academic achievement and the development of independent thinking remains a subject of ongoing scholarly debate. Contemporary studies point to the potential of AI-supported learning environments while simultaneously emphasizing the risks of diminished cognitive autonomy when intelligent systems are used without pedagogical regulation. The present study aims to examine the effectiveness of an AI-oriented approach to mathematics instruction based on the combination of personalized tutoring support and learning analytics tools within a university educational context. Within a quasi-experimental mixed-methods design, an AI-supported instructional model was implemented, incorporating adaptive learning pathways, an intelligent tutor providing step-by-step scaffolding, and a system for analysing students’ learning behaviours. Participants were undergraduate students enrolled in university-level mathematics courses. The effectiveness of the approach was assessed using pre- and post-test results, log data capturing students’ interactions with the AI system, and indicators of academic engagement and independent problem-solving. The results demonstrate that the use of personalized AI-based tutoring is statistically significantly associated with improved academic performance, the development of independent problem-solving strategies, and a reduction in the proportion of students classified as academically at risk. Furthermore, learning analytics data analysis enabled the identification of behavioural indicators with predictive value for the early detection of learning difficulties. Overall, the findings confirm that, when integrated within a sound pedagogical framework, artificial intelligence does not replace students’ cognitive activity but rather functions as a supportive tool for fostering independent mathematical thinking and for informing evidence-based instructional decision-making. The practical significance of this study lies in the proposal of a reproducible AI-oriented model for mathematics education that can be applied across university courses of varying levels of complexity..

Keywords: Artificial intelligence in education, university mathematics education, personalized learning, intelligent tutoring systems, learning analytics, independent mathematical thinking, academic achievement, at-risk students.

Full Text PDF

This article is published under the terms of the Creative Commons Attribution License 4.0