The role of communication in health-focused conversational AI: a review of systematic reviews examining conversational features, design characteristics, and interaction functions
Sue Lim, Qinghua Yang, Jingbo Meng, Stephen A. Rains · Behaviour and Information Technology · 2025
Conversational artificial intelligence (AI) involving software capable of engaging in human-like interaction has been increasingly employed as part of interventions to address a wide range of health conditions. Despite the central role of communication in health-focused conversational AI, the communication characteristics of these technologies have not received the attention they warrant. To fill this gap in the literature, we systematically reviewed 79 systematic or meta-analytic reviews of health-focused conversational AI from a communication perspective. We leveraged the Computers are Social Actors (CASA) framework as a guide to examine three dimensions of health-focused conversational AIs: (a) conversational features, (b) design characteristics, and (c) interaction functions. Modality was the most reviewed conversational feature, and conversational AIs’ appearance or anthropomorphic look received the most scholarly attention as a design characteristic. The majority of reviews examined social support as conversational AIs’ interaction function, followed by education and training/coaching. Overall, our project offers a comprehensive summary of the role of communication in health-focused conversational AI and identifies constructs and theories to prioritise in future scholarship.