<![CDATA[Purpose: The rapid expansion of artificial intelligence (AI) has reshaped mathematics teaching and learning by introducing adaptive support, automated feedback, and new forms of interaction. At the same time, its use has raised pedagogical, ethical, and institutional concerns. This study aims to examine the opportunities, risks, and pedagogical safeguards associated with integrating AI into mathematics education.Method: This study employed a Systematic Literature Review following the PRISMA 2020 guidelines. A systematic search was conducted in ERIC, Web of Science, and Scopus for peer-reviewed journal articles published between 2020 and 2025. From 309 identified records, 40 studies met the eligibility criteria and were included in the final synthesis. The data were analysed through descriptive mapping and thematic analysis involving familiarisation, initial coding, theme grouping, and interpretation.Findings: The review identified that AI supports personalised learning, automated assessment, immediate feedback, student engagement, independent practice, and instructional planning. Intelligent Tutoring Systems and adaptive learning platforms were the most frequently examined technologies. However, AI integration was also associated with unequal access, limited teacher readiness, weak pedagogical alignment, superficial learning, dependence on generated answers, and concerns about fairness and transparency. The synthesis further showed that responsible implementation requires teacher mediation, structured scaffolding, verification of AI-generated responses, reasoning-oriented tasks, learning-process monitoring, and human oversight.Significance: This review brings opportunities, risks, and pedagogical safeguards together within a single analytical framework. It positions AI as a pedagogical support system that should remain aligned with mathematical reasoning, ethical responsibility, and contextual needs]]>
Copyrights © 2026
