The emergence of generative artificial intelligence (GenAI), in particular, large language models (LLMs), is reshaping mathematics education by offering new forms of tutoring, feedback, problem generation, and personalized learning. This systematic review, following the PRISMA methodology, synthesizes empirical studies published from 2020 to 2025 that examine the application of GenAI in mathematics learning at secondary and tertiary levels. From an initial yield of 68 records across multiple databases, 32 were retained after screening and quality assessment, of which 22 met criteria for detailed synthesis. The review finds that GenAI tools can significantly support mathematical problem solving, provide timely feedback, adapt instruction to learner needs, and increase accessibility — particularly for learners with special educational needs. Nonetheless, critical challenges persist, including risks of hallucinated or incorrect solutions, student over-reliance that weakens procedural fluency, and significant academic integrity challenges. The field also lacks long-term evidence on conceptual mastery. The paper concludes that GenAI must function as a scaffolding tool rather than a solution generator, necessitating robust teacher training, resilient assessment redesign against misuse, and prioritizing longitudinal and large-scale studies.

Authors:
Jennifer Dela Torre, Higher Colleges of Technology, United Arab Emirates
Jero Sayco, Higher Colleges of Technology, United Arab Emirates

About the Presenter(s)
Dr. Jennifer Dela Torre is a Mathematics Lecturer at HCT, UAE, interested in technology-enabled math learning and AI integration in education. She is currently developing AI-enhanced tools for personalized mathematics instruction.

See this presentation on the full schedule – On Demand Schedule

Bookmark