Drawing on compensatory control theory, this study aims to examine how AI recommendation agent type (autonomous AI vs. collaborative AI) influences tourists’ adoption intention. It further investigates the moderating role of tourists’ occupational type and the mediating role of perceived control.
Three scenario-based experiments were conducted. Experiment 1 tested the general effect of AI recommendation agent type on adoption intention. Experiment 2 examined the moderating role of occupational type (manual labor vs. mental labor) and the mediating role of perceived control in the relationship between AI recommendation agent type and adoption intention. Experiment 3 manipulated participants’ temporary control state, namely, control deprivation versus control affirmation, to provide causal evidence for the proposed control-based mechanism.
When tourists’ occupational type was not differentiated, tourists were generally more willing to adopt recommendations from autonomous AI than from collaborative AI. Occupational type significantly moderated this relationship. Manual laborers were more willing to adopt recommendations from collaborative AI, whereas mental laborers were more willing to adopt recommendations from autonomous AI. Perceived control in the AI recommendation interaction mediated the interaction effect of AI recommendation agent type and occupational type on adoption intention. Experiment 3 further showed that the temporary control state changed individuals’ preferences for autonomous versus collaborative AI.
This study extends tourism AI adoption research by shifting attention from AI-generated content and tourists’ general perceptions of AI technologies to the autonomy of AI recommendation agents. It also enriches compensatory control theory by applying it to AI-enabled tourism recommendation contexts and distinguishing different sources of perceived control provided by autonomous and collaborative AI.
