When Designers Sweat: Behavioral Traces of GenAI Co-Creation

The integration of generative artificial intelligence (GenAI) into design processes raises fundamental questions about behavioral patterns in human-GenAI interaction. This study examines how 16 professional designers interact with GenAI tools during concept development through a mixed-methods approach including pre/post-task questionnaires, video-based behavior analysis, and digital interaction tracking. Results reveal a critical distinction between reflective usage modes and creative modes, with differentiated cognitive impacts. Analysis of communication loops shows significant correlations between interaction difficulties and final design output quality. Three distinct clusters emerge: designers with fluid, problematic, and adaptive interaction patterns. By providing a methodological framework for evaluating GenAI tool effectiveness in design practice, this research contributes to theoretical understanding of behavioral processes in human-GenAI co-creation. Findings reveal specific strategies and workflow adaptations that optimize designer-GenAI collaboration, informing both design methodology and human-computer interaction practice.

• Human-centered computing → Empirical studies in HCI.