Designing conversational AI for digital healthcare in regards to accessibility

Abstract: The area of Conversational AI in healthcare is gaining traction, especially in this rapid and global spread of COVID-19. An increased number of digital services have emerged in order to tackle this pandemic. Conversational AI is an example found in this study that can be used to quickly get information about symptoms and help the individual in their self-assessment. Users can quickly and easily can get in touch with healthcare professionals in ways that work for the individual, who no longer has to rely solely on phone calls. In this study, a proof-of-concept is designed and evaluated together with three participants with cognitive disabilities and an experienced project manager. The method used is participatory design where participants were invited to contribute to the design process during four workshops. The results show that conversational AI can make healthcare more accessible for people with disabilities, which is confirmed by the literature as well as in the user tests conducted. The paper results in four important factors when designing Conversational AI in digital healthcare: Be consistent, use clear and understandable language, provide alternative ways of interaction and test with real users.