Trust and distrust in ChatGPT: a configurational perspective on tourism information adoption

Purpose This study aims to explore how trust and distrust toward generative artificial intelligence (AI), particularly ChatGPT, are formed and how they jointly shape the adoption of AI-recommended travel information. Moving beyond prior research centered on technical attributes, it conceptualizes trust and distrust as coexisting and asymmetric mechanisms in tourism technology adoption. Design/methodology/approach Drawing on the trust and distrust duality framework, the study models information usefulness, perceived intelligence and competence as antecedents of trust, and hallucination anxiety and privacy concern as drivers of distrust. Data were collected from 473 users of ChatGPT for travel information. Analyses used structural equation modeling, mediation testing, moderated regression and a complementary response surface-based analysis to further examine the joint, asymmetric and configurational effects of trust and distrust on perceived value. Findings Trust enhances perceived value and adoption intention and mediates the impact of enabling factors. Hallucination anxiety and privacy concern increase distrust, which does not exert a direct effect on perceived value but operates as a conditional and suppressing mechanism that shapes the impact of trust. Personal innovativeness negatively moderates the distrust–value link. Originality/value This study advances the literature by moving beyond treating trust and distrust as parallel or opposing predictors, demonstrating that they operate as configurational and asymmetric mechanisms whose effects emerge through their joint and conditional interplay. By doing so, it offers a novel perspective on value formation in AI-mediated tourism contexts and extends existing technology adoption frameworks.