Florence Chatbot Evaluations, WHO Dialogic Commitment, and Vaccination Intention: The Conditional Role of Misinformation Involvement

Background/Objectives: Public-health chatbots convey information while representing health organizations. We examined whether evaluations of WHO’s (World Health Organization) Florence chatbot were associated with WHO Dialogic Commitment and Vaccination Intention, and whether misinformation Involvement conditioned the indirect association of Chatbot-Use Risk Perception with Vaccination Intention. Methods: We analyzed a cross-sectional survey of 416 complete-case U.S. respondents with prior chatbot experience who interacted with Florence. Confirmatory factor analysis and demographic-adjusted latent-variable structural equation modeling assessed the main associations. A standardized-composite moderated-mediation model used 5000 bootstrap samples. Latent product-indicator, vaccination-history-adjusted, and full-information maximum-likelihood models provided sensitivity checks. Results: The six-factor measurement model fit well, χ2(155) = 316.25, CFI = 0.973, RMSEA = 0.050, SRMR = 0.051. Chatbot Competency and Emotional Connection were positively associated with WHO Dialogic Commitment, which was positively associated with Vaccination Intention. The average Chatbot-Use Risk Perception path was nonsignificant. Its interaction with Involvement was positive (b = 0.079, 95% bootstrap CI [0.023, 0.133]); the indirect association through WHO Dialogic Commitment was negative at low and mean Involvement but not detectable at high Involvement. A latent-interaction sensitivity model yielded a similarly positive interaction. Conclusions: Respondents’ Florence evaluations covaried with perceived WHO dialogue and vaccination-related intentions. Involvement attenuated the negative risk–dialogue association. Mixed initial-vaccination and booster meanings, nonprobability recruitment, and same-session self-report limit causal and population-level interpretations.