Mathematical Problem Posing and Generative Artificial Intelligence: anexperience in a public elementary school

Abstract

This article examines how Mathematical Problem Posing, mediated by Generative Artificial Intelligence, is configured in the teaching and learning of mathematics in an elementary school class at a public school in Northeastern Brazil. The study adopts a qualitative approach and is designed as Pedagogical Research, grounded in the theoretical construct of humans-with-media/AI and contemporary research on Problem Posing. The empirical experience involved the use of ChatGPT to explore first-degree equations with one unknown, drawing on students’ written records, classroom interactions, and AI-mediated dialogues as data sources. The analysis of narratives indicates that, while generative AI can support initial problem exploration and strategy formulation, its pedagogical contributions are strongly conditioned by students’ mathematical literacy, reading comprehension, material conditions, and teacher mediation. The findings highlight that the integration of generative AI into Problem Posing requires pedagogical intentionality, specific teacher education, and digital inclusion policies to become an effective formative practice.