AI-driven personalization is the key technology for tourism platforms and influences people’s behavior when searching, evaluating, and choosing experiences in the post-pandemic period. Though AI personalization assists travelers in coping with information overload and improves fit of options, it leads to increased privacy and manipulation problems due to data-intensive and black box-like inference of algorithmic decision-making (Awad and Krishnan, 2006; Aguirre et al. 2015; Lei et al. 2022). Following the Theory of Planned Behavior (TPB), the research develops and empirically tests an extended TPB framework that adds trust in AI personalization, perceived value, and perceived risk as the determinants of attitude and intention and ties intention to usage and decision quality (Ajzen, 1991; Hair et al. 2022). With the help of a survey of tourists exposed to AI-powered personalization services on tourism platforms (n = 420), the proposed model is tested using PLS-structural equation modeling (PLS-SEM) that involves measurement reliability and validity, HTMT discriminant validity, common method bias tests based on collinearity analysis, and bootstrapping of hypotheses (Henseler et al. 2015; Kock, 2015). Attitude, subjective norms, and perceived behavioral control were revealed to be the important predictors of intention. Trust in AI and perceived value were recognized as predictors of attitude and intention while perceived risk was the determinant of their reduction. Besides, intention predicted personalization usage and both usage and intention influenced positively perceived decision quality. Hence, the research contributes to the extension of TPB in the case of AI-personalized tourism by means of theoretical justification of trust-value-risk framework and privacy calculus.
