AI Tourism Service-Use Breadth and Perceived-Benefit Breadth

Abstract

Aims: This study examines whether broader use of AI-enabled tourism service categories is associated with recognition of a wider range of perceived benefits after age and travel frequency are taken into account. Study Design: A cross-sectional secondary analysis was conducted using a predefined data-collection window from an ongoing online AI tourism survey project. Place and Duration of Study: The survey was administered online through Tencent Questionnaire. Responses collected between April 1 and June 27, 2026, were included in the present analysis. Methodology: The analytical sample comprised 496 valid responses. Two formative category-count measures were constructed: the AI Use Index (0–5), representing AI tourism service-use breadth, and the Positive Index (0–4), representing perceived-benefit breadth. Hierarchical linear regression was used as the primary analytical approach, with age and travel frequency included as control variables. Ordered logistic regression and sensitivity analyses using an alternative benefit measure and the exclusion of contradictory multiple-choice responses were conducted to assess the stability of the main association. Results: Adding the AI Use Index increased the explained variance in the Positive Index from 7.6% to 33.7% (ΔR² = .261). AI service-use breadth was positively associated with perceived-benefit breadth after age and travel frequency were taken into account (B = 0.481, 95% CI [0.413, 0.549], P < .001). The ordered logistic model produced a consistent result (OR = 3.836, 95% CI [3.107, 4.735], P < .001). The association remained significant after excluding the provider-oriented reduced-labour-cost category and after removing contradictory multiple-choice responses. Conclusion: Respondents who reported experience with a broader range of AI tourism service categories also tended to recognise a broader range of potential benefits. Service-use breadth therefore provides a useful dimension for examining increasingly multifunctional AI tourism environments. Because the study used cross-sectional, self-reported, non-probability survey data, the findings should be interpreted as associations rather than causal effects or population estimates.