This study implemented a generative AI-based interactive questionnaire system and exploratorily examined its potential in supporting the verbalization of tacit knowledge. Thirteen art and design students participated in a chatbot-style survey examining their perceptions of white space in visual design. The system utilized GPT-4.1-nano and aimed to realize semi-structured interview characteristics through contextually adaptive questioning. Results indicated that 77% of participants felt they could effectively articulate their thoughts through dialogue, with positive evaluations regarding empathetic responses and reduced perceived burden despite 19 questions. However, challenges emerged in dialogue naturalness (38% positive) and thought deepening (46% positive). These findings suggest that implementing sophisticated probing question generation based on semantic analysis is essential for achieving fully functional semi-structured interviews. The system was able to elicit diverse perceptions about white space, including physical versus meaning-based interpretations. This exploratory study provides preliminary insights into the characteristics, potential, and limitations of AI-based interactive questionnaires.
OTSUE et al. (Thu,) studied this question.