Integration of Language and Visual AI Models in Kansei Engineering

Sammanfattning

Kansei Engineering translates emotional responses into product designs, traditionally requiring significant human effort. With Artificial Intelligence (AI) emerging as a powerful tool, this study examines AI's potential integration within the Kansei Engineering framework to enhance traditional human-centered design approaches. We evaluated different AI models across distinct methodological stages by implementing a two-phase research approach using mechanical keyboards as our case study. First, we conducted a complete Kansei Engineering process following the traditional methodology. Second, we performed the same steps while incorporating large language models and generative image models, followed by a systematic comparative analysis between these approaches. Our results revealed that AI can effectively generate affective product images, process data accurately, and produce relevant Kansei words that expand solution spaces. AI demonstrated strengths in establishing target group understanding through persona creation aligned with survey data, and in efficiently classifying and restructuring information. Notably, AI-generated keyboard images sometimes received higher ratings than real product images in conveying desired impressions, though with some limitations in product property identification. However, this evaluation relies on a manually executed Kansei Engineering method as the baseline, which may contain inherent limitations that could influence AI performance assessments. Additionally, findings are specific to keyboards and may vary across different product categories. This research demonstrates that AI integration is most effective when supporting time-intensive methodological tasks rather than replacing expert judgment. The technology enhances efficiency in data handling and preliminary exploration while providing visualization capabilities that offer valuable alternatives when speed and iteration are prioritized over perfect fidelity. These findings contribute to understanding how emerging technologies can complement traditional design approaches, potentially improving product development when strategically implemented in specific stages of Kansei Engineering.