#200: Patterns of Student-AI Collaboration in Problem-Solving: An Ordered Network Analysis

This research investigates interaction dynamics between graduate students and AI agents during collaborative problem-solving activities within a curriculum design context. Analyzing 1,569 conversational exchanges, we applied ordered network analysis to map behavioral sequences and identify emergent patterns across problem-solving phases. Three dominant interaction pathways emerged: students soliciting AI perspectives, followed by resource provision; students seeking clarification with AI, which responds to refine their understanding; and students prompting elaboration that triggers extended cognitive exchanges. These patterns reveal AI functioning beyond a passive information source, instead operating as an active collaborator that reciprocally shapes problem-solving discourse. Findings indicate the formation of hybrid intelligence environments where human-AI partnerships distribute cognitive work. This research contributes theoretical insights into human-AI collaboration, demonstrates ordered network analysis as a methodology for examining conversational patterns, and offers practical implications for designing educationally productive AI-mediated learning experiences.