Students' perspectives on GPT-4's supported problem solving

In this paper we present an exploratory study aimed at investigating upper secondary school students' perspectives on the role that GPT-4 could play in scaffolding their problem solving processes. Data have been collected by means of semi-structured interviews developed before, during and after students' interaction with GPT-4. The data analysis, developed by the IPA methodology, is presented by focusing on a paradigmatic example of a 15-year-old student. The results of the analysis highlight clusters of themes that allow us to shed light on the complexity of the instrumental genesis processes that characterize human-Large Language Models (LLM) interaction, suggesting that this complexity should be taken into account in the design of suitable tasks involving LLM.