A risk-benefit account of AI tourism adoption: prior usage, familiarity, usefulness, and AI misinformation concern in travel planning

Generative artificial intelligence (AI) tools such as ChatGPT are increasingly used for tourism planning, but they can also produce false or misleading information. This study uses a risk-benefit account to examine how prior AI usage, AI familiarity, perceived ease of use, perceived usefulness, and attitude (the benefit side) combine with AI misinformation/error concern (the risk side) to shape behavioral intention to use AI for tourism planning. Based on public survey data from 900 consumers, the study estimates a confirmatory factor analysis (CFA) measurement model and a structural equation model (SEM) with robust maximum likelihood and full-information maximum likelihood (FIML). The main model, grounded in the Technology Acceptance Model (TAM) and Theory of Planned Behavior (TPB), shows a strong benefit pathway: prior usage predicts familiarity, familiarity predicts ease of use, ease of use predicts usefulness and attitude, and both usefulness and attitude predict intention. Familiarity is also positively associated with AI misinformation/error concern, indicating that more familiar users are more aware of possible AI errors. However, that concern has no significant direct association with intention, and robustness checks do not support any moderation by concern. The study contributes a risk-benefit account in which AI error awareness coexists with positive adoption beliefs but does not outweigh them in this lower-stakes, verifiable tourism-planning context.