Using SOR and Psychological Reactance Theory, the study proposes a Threat-Substitution Mechanism for high-involvement online travel decisions. PLS-SEM on 488 Chinese Gen Y/Z OTA users found AI personalization increased usage intention and reduced reactance, with partial mediation. Privacy concern strengthened the negative AIP-reactance relationship, interpreted as heightened threat-appraisal sensitivity rather than uniform resistance.
Key findings
- AIP increased intention and reduced reactance; reactance partially mediated; privacy concern moderated the AIP-reactance path; utilitarian benefits dominated while friction remained.
Why this matters globally
Platforms should combine useful personalization with autonomy controls, explanation and privacy choices rather than hidden tracking, because durable adoption depends on utility and control.
Thai researcher contribution
Hongmei Duan and Ahmad Yahya Dawod are affiliated with Chiang Mai University in cross-country AI commerce and user-behavior research.
Limitations to consider
Chinese Gen Y/Z OTA users do not represent other populations. Cross-sectional PLS-SEM does not establish causality or actual continuance. Construct wording and common-method variance may shape paths, and the counterintuitive moderation needs behavioral replication.