Enhancing Explainability in Medical AI: Developing Human-Centered Participatory Design Cards

Explainable artificial intelligence (XAI) aims to develop AI systems that are easy for humans to understand and explain the decision-making process clearly. Especially in high-risk fields such as medical, the explainability of AI systems is particularly crucial. However, designers face significant challenges in implementing explainable AI design activities in the medical domain. This research aims to provide systematic guidance for designers on medical XAI design methods, addressing the challenges faced by designers and users in designing explainable medical AI. Through comprehensive literature review and thematic analysis based on the PRISMA process, we developed medical XAI Design cards from a human-centered perspective to assist designers in exploring solutions. Combining qualitative and quantitative research, we collected feedback from designers and users, validating the effectiveness of the Design Cards and aiding in maintaining a focus on medical explainability during the design process. This research fills the gap in enabling designers to engage in medical XAI design activities, enhancing the explainability of XAI systems in the medical field.