Evaluating AI and Teacher Explanations in High School Geometry to Support Sustainable Education

While AI chatbots are increasingly used by high school students when solving curriculum-based geometry problems, empirical evaluation of their perceived effectiveness has not been explored sufficiently. This paper studies students’ experiences of using ChatGPT for geometry tasks and compares AI explanations with those provided by high-school teachers. A 14-question survey of 102 Grade 11 students in Nazarbayev Intellectual School of Science and Mathematics in Karagandy has been conducted to analyze student perceptions of the use of AI, its accuracy, errors, and hallucinations. The findings demonstrate that while ChatGPT was frequently correct in final answers, its low effectiveness in long-term geometry skill acquisition and apparent hallucinations in 3D geometry were reported. To summarize, the study brings valuable Kazakhstani experience of AI usage for mathematics to the global research on artificial intelligence integration in education. This research paper has been awarded second place at the V Annual Global Perspectives, Project Work, Economics and MUN Conference held in Astana (Nazarbayev Intellectual School of Science and Mathematics in Nura district of Astana) on April 25, 2026.