An Exploratory Study on Fashion Design Ideation Experiences and Creativity Using Generative AI-Based Visual Stimuli

The rapid development of generative AI technologies is creating new opportunities for design ideation and human-AI co-creation (HAIC) in the fashion industry. This study evaluated experiences associated with generating and utilizing generative AI-based visual stimuli in fashion design ideation and examined the characteristics of resulting creative outcomes. An exploratory mixed-methods approach was employed with 10 fashion design students who completed the same design task using Midjourney and Pinterest. Data were collected through design experiments, surveys, in-depth interviews, and expert evaluations. The findings indicated that generative AI-based visual stimuli functioned as interactive resources supporting cognitive exploration during ideation. Participants perceived generative AI as a useful tool for quick visualization and idea exploration, enabling the visual development of abstract concepts through simple prompt inputs. Participants also engaged in iterative processes of refining prompts, transforming images, and regenerating design ideas. Furthermore, survey and interview results showed positive evaluations for usability and perceived value, whereas predictability, communication satisfaction, and sense of accomplishment received lower ratings. Repetition and unpredictability in output emerged as key characteristics of the generative AI experience. Some participants regarded these features as obstacles, while others viewed them as opportunities for unexpected visual combinations and new design possibilities. Expert evaluations showed higher levels of design fixation in the Pinterest condition, whereas the generative AI condition demonstrated lower fixation levels and several highly creative cases. However, the median creativity scores were identical across conditions, suggesting that the creative advantages of generative AI should not be generalized. Additionally, differences between expert evaluations and participants’ perceptions of creativity indicate that evaluative perspectives can lead to varying interpretations of creative outcomes. Overall, the findings indicate that generative AI and online curation platforms support fashion design ideation in different ways. Generative AI can complement existing visual reference resources by providing a supportive environment for iterative exploration during ideation. The findings highlight the importance of considering user experience alongside design outcomes in HAIC contexts.