Modeling the Creative Thinking Processes of Design Experts and Novices

Abstract

The rapid advancement of generative AI is transforming design workflows from manual "modeling" into the sophisticated "selection" and "reconstruction" of a multitude of generated options. Within this paradigm shift, it is a pressing issue to clarify the boundary between "implementation capabilities," which AI can rapidly replace, and the essential "curation capabilities" that humans must assume. This research aims to establish a definition of the new creative thinking processes required in the AI era from an ontological perspective by visualizing the unconscious cognitive processes of experienced designers. Previous research on design expertise has faced difficulties in extracting pure cognitive structures because physical factors, such as drawing skills, act as noise. Therefore, this research verified the differences in cognitive strategies between experts and novices using a unique abstract composition task that eliminated physical skills and tool familiarity. Experiments were conducted with 10 experts and 10 novices, and an integrated analysis of operation logs and interviews using the FBS (Function-Behavior-Structure) framework revealed that experts spent approximately twice as much time on "pause" and "evaluation" processes compared to novices. These results demonstrate that while the novice's creative process is a "linear task of filling in blanks," the expert's process is a "recursive exploration" involving repeated critical dialogue with the generated results based on internal criteria. This "intentional pause" constitutes the tacit knowledge unique to humans in the AI era and lies at the core of future design education. Ultimately, this suggests that the role that humans must play is shifting from simple implementation to advanced curation and judgment capabilities.