An Investigation of ChatGPT–Student Interaction ın The Learning of Functions

Abstract

This study aims to examine the cognitive orientations of prompts generated by 10th-grade students in their written interactions with ChatGPT while learning the topic of functions, as well as the explanatory strategies employed in ChatGPT’s responses. The research focuses on describing interaction patterns and the emerging learning needs. Conducted within a qualitative approach using an exploratory case study design, the study involved eight students (5 female, 3 male) enrolled in a private science high school located in the southern region of Türkiye. The primary data source consisted of student–ChatGPT prompt–response transcripts recorded during the implementation process; a worksheet on functions was used to structure the learning context. The data were analyzed through content analysis using open coding, and the prompts were classified under five categories. The findings indicated that the prompts were predominantly Definition-Oriented (41.2%) and Representation-Oriented (27.5%), while ChatGPT’s responses were distributed as Exemplifying (35.3%), Visualizing (33.1%), and Preventive (31.6%). The results revealed that ChatGPT has the potential to support conceptual clarification and representational enrichment; however, learning outcomes were found to be associated with the quality of students’ questioning as well as their verification and self-regulation behaviors. Based on these findings, several recommendations were presented regarding the effective use of generative artificial intelligence tools in mathematics education, particularly emphasizing pedagogically structured usage scenarios and guidance strategies.