Deep thinking function in AI-mediated travel planning: how reasoning transparency shapes information adoption and destination acceptance

Abstract
This study investigates how DeepSeek’s Deep Thinking Function, which introduces reasoning transparency into AI-mediated travel planning, shapes travelers’ psychological evaluation and adoption of AI-generated destination recommendations. Drawing on an extended Information Adoption Model (IAM), the study conceptualizes Deep Thinking Function not merely as a system design feature, but as a cognitive persuasion mechanism that reconstructs users’ judgment criteria through process-based reasoning disclosure. A purposive sampling strategy was applied to recruit tourism users who had recently interacted with AI travel assistants, yielding 260 valid responses collected via an online survey. Partial Least Squares Structural Equation Modeling (PLS-SEM) was employed to assess the research model. The results reveal that Deep Thinking Function significantly enhances perceived argument quality and source credibility, which subsequently strengthen perceived information usefulness, information adoption, and destination acceptance intention. The findings indicate that, in AI-generated information environments, travelers’ adoption decisions shift from outcome-focused evaluation toward process-oriented cognitive appraisal, where analytical competence and methodological rigor become core foundations of trust. This study extends IAM into AI-mediated tourism contexts, conceptualizes transparency as an active persuasion mechanism rather than a static ethical attribute, and demonstrates the emergence of system centric cognitive trust in travel decision-making. Practical implications highlight the value of process aware AI interface design, transparency driven persuasion strategies, and the repositioning of destination branding toward decision logic compatibility in post-AIGC tourism ecosystems.