ChatGPT 기반 디자인 과정 분석: 디자인 사고 마인드셋 수준에 따른 차이를 중심으로

초록

This study examines how learners interact with ChatGPT across design activities when performing a design task and how these patterns vary by levels of design thinking mindset in a generative AI–based design learning context. Sixteen undergraduate students from a university in Seoul, organized into four groups of four, completed an educational chatbot design task using ChatGPT. Interaction logs were coded using Safoutin et al.’s (2000) design attributes framework across ten categories: need recognition, problem definition, planning, information gathering, idea generation, modeling, feasibility analysis, evaluation, implementation, and documentation. Design thinking mindset scores were used to divide teams into high- and low-mindset groups, and Ordered Network Analysis (ONA) was used to examine transitions among design activities. Results showed that modeling functioned as a central hub, forming recurring transition patterns across activities. Transitions centered on implementation, linking idea generation, planning, and evaluation, were more prominent in the high-mindset group, whereas transitions centered on modeling and problem definition were more salient in the low-mindset group. These findings suggest that even within the same AI-based learning environment, learners’ use of ChatGPT and the unfolding of design processes may differ by mindset level. The results further indicate that ChatGPT can function as an interactional resource connecting diverse design activities throughout the learning process, providing implications for differentiated instructional design and AI-supported pedagogical strategies in generative AI–based design education.

키워드