Asymmetric trust mediation of perceived intelligence and explainability in AI-powered tourism platforms

Purpose This study aims to examine how perceived intelligence and perceived explainability, as two theoretically distinct AI-related perceptions, are associated with consumer trust and purchase intention in tourism platforms characterized by algorithmic opacity and high-stakes consumption decisions. Design/methodology/approach Drawing on signaling theory and the Stimulus-Organism-Response framework, with trust interpreted through an Ability-Benevolence-Integrity-informed conceptual lens, this study develops and tests a signal-differentiation framework using PLS-SEM with data from 240 Chinese tourism platform users. Findings The findings reveal a signal-level asymmetry in how AI-related perceptions are associated with purchase intention. Perceived intelligence is associated with purchase intention both directly and indirectly through trust, although the direct association is comparatively modest and warrants conservative interpretation, whereas perceived explainability is associated with purchase intention through trust rather than through a significant independent direct pathway in the present model. This pattern is consistent with the interpretation that competence-related perceptions are linked to both utility-related and trust-related evaluations. In contrast, transparency-related perceptions show behavioral relevance mainly through trust-related evaluation. Exploratory moderation analysis further shows education-related heterogeneity in the explainability–trust relationship, whereas the intelligence–trust relationship remains comparatively stable across education levels. Given the cross-sectional design, the results should be interpreted as association patterns rather than causal evidence. Originality/value This study advances a signal-differentiation perspective by showing that AI-related perceptions differ not only in content but also in the type of information asymmetry they address, with competence-oriented perceptions linked to adverse-selection-type uncertainty and transparency-related perceptions linked to moral-hazard-type uncertainty, resulting in structurally distinct trust-related pathways to purchase intention.