Seoul Haeseong Primary School, Seoul, South Korea

Abstract: Educational digital transformation has become a strategic direction of global basic education reform, and artificial intelligence technology is profoundly reshaping the teaching ecology of mathematics classrooms. Taking primary mathematics teaching in South Korea as the research background, this study systematically explores the theoretical basis, practical paths, and rational boundaries of AI-enabled primary mathematics teaching, combining the national policy practice of introducing AI digital textbooks into Grade 3 and Grade 4 primary mathematics classrooms in South Korea since 2025. This paper proposes a three-in-one AI teaching application framework of “intelligent diagnosis, accurate supply, and dynamic evaluation”, and conducts in-depth discussion from three dimensions: the construction of personalized learning paths, the reconstruction of situational teaching scenarios, and the upgrading of teaching evaluation systems. The research finds that AI technology can effectively resolve three core contradictions in traditional mathematics teaching: the conflict between unified teaching progress and individual differences, the separation between abstract concepts and concrete understanding, and the imbalance between result-oriented evaluation and process-oriented attention. However, the application of technology must adhere to the basic principle of “assistance rather than substitution”, taking the development of students’ mathematical thinking as the fundamental goal. Combined with front-line teaching cases, this study provides operable practical paths for primary mathematics teachers to realize the role transformation from “technology users” to “intelligent guides” in the digital era.

Keywords: artificial intelligence; primary mathematics; educational digitalization; personalized learning; teaching evaluation