Affectivegen: an exploring methodology for emotional design implementation

Emotional design has become a cornerstone of contemporary design, offering a key framework for addressing user needs and fostering innovation. However, although generative models demonstrate strong content creation capabilities, their application in emotional design remains limited. Existing models often provide insufficient support because their processes are opaque and their outputs overly simplistic, making them less suitable for complex emotional design tasks. This study proposes a methodology, integrating generative models with the AEIOU method (Activity, Environment, Interaction, Object, User), a pioneering model and Kansei engineering principles to support emotional design ideation. The integration enables multi-step reasoning and structured ideation, supporting more emotionally resonant design exploration. To implement this approach, we developed AffectiveGen, an interactive system with a mind map interface that visualises reasoning chains and enables iterative refinement of generative workflows through input and output control. Two user studies demonstrated that AffectiveGen significantly improved design quality and emotional expressiveness while strengthening human-AI collaboration by helping designers articulate and refine creative intent. This research establishes a theoretical and methodological foundation for future emotional design applications of generative models. It also highlights practical pathways for integrating emotional intelligence mechanisms into large language model-driven design systems for real-world creative practice contexts.