Investigating the Division of Labour in Student-AI-Collaboration on Critical Thinking Tasks

This study investigates the division of labor between students and generative AI in solving tasks addressing critical thinking. Three key research interests guide the study: (1) the extent of ChatGPT usage by students, (2) the division of labour between student and AI as perceived by students, ChatGPT, and an expert, and (3) if these ratings align with each other. An experiment involving 20 undergraduate students who completed 11 tasks using ChatGPT was carried out. Findings demonstrated substantial variation in the extent of ChatGPT usage among students. The division of labour analysis revealed that ChatGPT is used most heavily for reasoning tasks, and that students consistently rated the AI's contribution to their task solutions as slightly lower than both ChatGPT and the expert. Correlations of the ratings indicated strong alignment between students and ChatGPT, but weaker alignment with the expert perspective, particularly in the metacognition domain. These results highlight the need for enhanced integration of AI in educational practices, focusing on fostering students' understanding of AI capabilities while promoting critical engagement. Future research should explore diverse populations, a broader range of task types, and the long-term impact of AI integration to enhance the effective use of generative AI in education.