Background of study: The rapid integration of Artificial Intelligence (AI) in higher education has transformed mathematical problem-solving practices, yet it introduces potential discrepancies in students’ visual understanding, particularly in geometric contexts, leading to commognitive conflicts in visual mediators. Aims and scope of paper: This study aims to analyze and compare the sources of commognitive conflict in iconic visual representations when students solve flat shape problems using conventional methods versus AI-assisted approaches. Methods: A qualitative design was employed involving 20 first-year mathematics education students divided into two groups (AI-assisted and conventional). Data were collected through students’ written work and semi-structured interviews and analyzed using the commognitive framework focusing on visual mediator components, with method and data source triangulation ensuring validity. Result: The findings reveal that while both approaches produce correct numerical solutions, conventional methods generate more accurate and coherent visual representations. In contrast, AI-assisted solutions tend to produce flawed or inconsistent diagrams that violate spatial constraints, triggering commognitive conflict in the iconic visual mediator (VM-I), and causing confusion among students. Conclusion: Conventional problem-solving remains more reliable for accurate visualization in geometry tasks, whereas AI use, although efficient, requires critical verification. The study highlights the necessity of integrating metacognitive checking and visual validation when using AI to minimize commognitive conflicts in mathematical learning.
Emanuel et al. (Thu,) studied this question.