Artificial Intelligence (AI)-based Chatbots in Promoting Health Behavioral Changes: A Systematic Review

Background Artificial-Intelligence (AI)-based chatbots can offer personalized, engaging, and on-demand health-promotion interventions. This systematic review evaluates the feasibility, efficacy, and intervention characteristics of AI-chatbots in promoting health-behavior change. Methods A comprehensive search was conducted in seven bibliographic databases (PubMed, IEEE Xplore, ACM Digital Library, PsychoINFO, Web of Science, EMBASE, and JMIR publications) for empirical articles published from 1980 to 2022 that evaluated feasibility and/or efficacy of AI-chatbots for behavioral change. The screening, extraction, and analyses of identified articles followed the PRISMA guidelines. Results Of the 15 included studies, majority studies ( n =11) reported high usability, acceptability and engagement, and some evidence on feasibility of AI-chatbots. Selected studies demonstrated high efficacy in promoting healthy lifestyles ( n =6), smoking cessation ( n =4), treatment/medication adherence ( n =2), and reduction in substance misuse ( n =1). Behavioral change theories and/or expert consultation were used to develop behavioral change strategies of AI-chatbots, including goal setting, monitoring, real-time reinforcement/feedback, and on- demand support. Real-time user-chatbot interaction data, such as user preferences and behavioral performance, were collected on the chatbot platform to identify ways of providing personalized services. The AI-chatbots demonstrated potential for scalability by deployment through accessible devices and platforms (e.g., smartphones and messenger). Participants also reported that AI-chatbots offered a non-judgmental space for communicating sensitive information. However, the reported results need to be interpreted with caution because of moderate to high risk of internal validity, insufficient description of AI-techniques, and limitation for generalizability. Conclusion AI-chatbots have demonstrated efficacy of health-behavior change interventions among large and diverse population; however, future studies need to adopt robust RCTs to establish definitive conclusions.