Lessons from Using ChatGPT in Calculus: Insights from Two Contrasting Cases

Abstract The integration of generative artificial intelligence (GenAI) is increasingly recognized for its potential to enhance mathematical learning and teaching. Although existing research on student perceptions of using ChatGPT in mathematical learning indicates a high intention to adopt the tool, little is known about how these perceptions influence students’ actual use of it in undergraduate-level mathematics. In this research project, students participated in a semester-long study that included six lab assignments designed to provide a formative learning experience, where they used ChatGPT to deepen their understanding of calculus concepts. This case study used a qualitative approach, following two participants among 12 students who had contrasting perceptions of ChatGPT’s role in mathematical learning. By conducting in-depth analyses of surveys, lab assignments, and chat histories, the study aimed to gain a deeper understanding of how these students applied ChatGPT to support their calculus learning. The results showed that students’ perceptions strongly impact their use of ChatGPT for learning. The learner with positive perceptions of ChatGPT engaged more frequently with the tool (actual usage), demonstrated a greater tolerance for ChatGPT’s mathematics errors (learner satisfaction), and showed greater motivation for its further uses (learning motivation). This usage pattern remained consistent despite the interventions implemented throughout the semester. The study concludes with formative pedagogical implications for integrating ChatGPT into mathematics education. Implications for mathematics teaching and directions for future research were discussed.