This study aims to explore how personalized services provided by pre-consultation chatbots influence patients' intentions to self-disclose and the boundary conditions of this effect. Drawing on the Communication Privacy Management Theory, it identifies AI interactive trust as a mediator and examines how disease sensitivity moderates this mechanism.
Three scenario experiments and a field survey were conducted to test hypotheses. Experiments 1 and 2 were between-subjects experiments that manipulated pre-consultation chatbot service personalization (low vs. high) in different disease contexts. Experiment 3 employed a 2 (service personalization: low vs. high) × 2 (disease type: low- vs. high-sensitivity) between-subjects design. Survey 4 reconfirmed all hypotheses.
Service personalization increased patients' willingness to self-disclose by enhancing AI interactive trust, but only for low-sensitivity diseases. For high-sensitivity conditions, this trust-based effect weakens or even disappears.
By uncovering the central role of relationship quality in human-AI interactions, this study delivers novel theoretical insights, distinct from prior benefit-risk frameworks, into the internal mechanisms by which AI service personalization shapes users' willingness to self-disclose, extending existing research on the mechanisms of personalized AI services.
