Autonomy Support and Agency in AI Health Chatbots: Motivational and Social Pathways in the Chinese Context

Abstract

As AI health chatbots become increasingly prevalent, the psychological mechanisms through which they influence user behavior remain insufficiently understood. Drawing on Self-Determination Theory and the Computers Are Social Actors paradigm, this study proposes and empirically tests a decomposed specification of the TIME-HAII action route, comprising two psychological pathways (a motivational pathway and a social pathway) within the Chinese sociocultural context. Data were collected and analyzed using a multi-method design integrating self-report measures and behavioral indicators. The findings show that the motivational pathway was robustly supported through indirect mechanisms, whereas the social pathway received only partial empirical support. Within the motivational pathway, perceived chatbot autonomy support was positively associated with users' perceived autonomy and self-efficacy, which sequentially mediated the relationship between autonomy support and behavioral alignment. Within the social pathway, chatbot agency facilitated interactional behavioral alignment through the activation of perceived presence. The findings advance theoretical understanding of human-AI interaction in health communication contexts and offer evidence-based design implications for AI-driven health systems.