Perceived AI Legitimacy and Trust in Human–AI Decision Assistance: A Dual-Path Perspective on Psychological and Normative Evaluation

As generative artificial intelligence (GenAI) increasingly supports decision-making, users evaluate not only the system’s performance and trustworthiness but also the legitimacy and appropriateness of AI decision processes. This study examines how Perceived AI Legitimacy (PAL) is associated with reliance-oriented decision-assistance intention through cognitive and affective trust, while examining the moderating role of personal need for structure (PNS). Survey data from 1,000 Korean adults with GenAI experience were analyzed using partial least squares structural equation modeling. PAL was positively associated with both trust dimensions and indirectly with intention through both, with a numerically larger indirect association through cognitive trust. PAL retained a small direct association, while PNS negatively moderated the affective trust–intention association. These findings indicate that willingness to incorporate AI-generated judgments reflects differentiated roles of legitimacy, competence-based confidence, and affective reassurance, with individual differences shaping how these factors relate to intention, although these associations should be interpreted as correlational.