Observer Perceptions of GenAI Use and Interpersonal Trust: The Roles of Warmth, Competence, and AI Literacy

Generative AI (GenAI) is widely used in learning and daily life, yet its associations with interpersonal trust remain unclear. Drawing on the Stereotype Content Model and social cognitive theory, this study (N = 406) examined whether perceived warmth and competence mediate the relationship between observers' perceptions of a target individual's GenAI use and interpersonal trust, and whether perceived AI literacy moderates this relationship. Results showed that GenAI use was negatively associated with interpersonal trust. Perceived warmth and competence statistically mediated this association: GenAI use was associated with lower perceived warmth and competence, which in turn were associated with lower interpersonal trust. Perceived AI literacy moderated the direct negative association. Specifically, this association was significant when perceived AI literacy was low but non-significant when high. However, perceived AI literacy did not moderate the indirect pathways. This study provides preliminary evidence for potential perceptual mechanisms and boundary conditions associated with GenAI use and interpersonal trust, offering insights for balancing efficiency and trust in human-AI collaboration.