Use of Artificial-Intelligence-Based Chatbots to Promote Physical Activity: A Systematic Review of Interventions

Background: Artificial-intelligence (AI) -powered chatbots have been increasingly utilized in interventions to promote physical activity and healthy lifestyle behaviors. This systematic review explored the effectiveness of these chatbots in physical activity interventions. Methods: A systematic review was conducted using comprehensive searches across PubMed, Web of Science, EBSCOhost, and the Cochrane Library. Studies included were peer-reviewed intervention studies utilizing AI-powered chatbots designed to promote physical activity, conducted from 2010 onward and published in English. Results: Of 11,347 articles identified, eight studies published between 2018 and 2023 met the inclusion criteria. Using randomized controlled trials or pre–post designs, these studies involved sample sizes from 15 to 116 participants. Interventions included chatbots, voice-activated assistants, and health-coaching applications in remote and controlled settings. Primary outcomes focused on physical activity metrics like moderate and vigorous physical activity and step counts, with secondary outcomes covering quality of life and engagement. Intervention effectiveness varied, with some showing significant improvements in physical activity while others had mixed results. Study quality assessment revealed generalizability, data collection, and missing data-handling limitations. Conclusion: AI-powered chatbots show promise as practical tools for promoting physical activity. They offer personalized, engaging, and scalable interventions that can be integrated into broader public health strategies. Despite their promise, the variability in effectiveness and inconsistencies in study quality underscore an urgent need for future research to standardize intervention designs and advance AI techniques for more reliable and scalable health outcomes.