The impact of health chatbot humanization on the proactive health behavior intention of users: an AI-interaction well-being paradox perspective

Although prior studies have shown that health chatbot humanization can foster proximal attitudinal and behavioral responses in individuals, little is known about whether and how it influences their more distal proactive health behavior intention (PHBI). Drawing on the technology–consumer well-being paradox model, this study examined the effect of health chatbot humanization on PHBI and its underlying psychological mechanisms, while also testing transparency as a boundary condition. A 2 (humanization: present vs. absent) × 2 (transparency: present vs. absent) between-subjects, agent-based interaction experiment (N = 403) was conducted in a simulated health consultation setting. Results show that humanization directly promoted PHBI, increased psychological closeness and AI-interaction well-being, and reduced perceived eeriness. Further analyses revealed that psychological closeness and AI-interaction well-being served as key mediators that link humanization to PHBI, whereas transparency weakened the positive effect of humanization on psychological closeness and the related indirect effects. By adopting a more comprehensive operationalization of humanization, this study extends our current understanding of the relational and affective mechanisms underlying the effect of humanization on PHBI, and it offers a more nuanced understanding of human-centered design in human–AI health communication.