Evaluation of primary school teachers' use and perceptions of artificial intelligence in primary school mathematics instruction: a mixed-methods study.

Introduction: This study examined primary school teachers' reported perceptions and self-reported uses of artificial intelligence (AI) in primary school mathematics instruction, with particular attention to how perceptions vary according to teachers' demographic characteristics and AI-related experiences. Methods: The study employed an explanatory sequential mixed-methods design and was conducted with 302 primary school teachers in Türkiye during the 2024-2025 academic year. Quantitative data were collected using the Scale for Assessing Teachers' Perceptions of Artificial Intelligence in Education. The quantitative findings were subsequently elaborated through semi-structured interviews with teachers who had prior AI-related training. Results: Teachers reported moderately positive overall perceptions of AI in mathematics instruction. Perception scores varied according to gender, professional experience, educational background, frequency of AI use, and prior AI training, although these differences were not consistent across all dimensions and some were practically modest. More frequent AI use and prior AI-related training were associated with more positive perceptions. Qualitative findings showed that teachers primarily positioned AI as a supportive resource for lesson planning, content and material development, visualization, problem solving, formative assessment, feedback, and out-of-class learning. ChatGPT, GeoGebra, and Photomath were among the most frequently mentioned tools. Teachers also identified concerns regarding age appropriateness, technological infrastructure, data privacy, reliability, teacher competence, and students' possible over-reliance on ready-made responses. Discussion: The findings indicate a generally favorable but cautious orientation toward AI in primary mathematics education. Teachers' willingness to use AI and positive attitudes do not necessarily reflect extensive or pedagogically established classroom integration. Rather than demonstrating instructional effectiveness, the study provides a context-specific account of the perceived pedagogical opportunities, practical constraints, and conditions associated with responsible AI integration in primary mathematics instruction.