The Surprising Paradox of AI Literacy: How Lower AI Literacy Can Lead to Higher Acceptance

This study examines the paradoxical relationship between consumers’ AI literacy and AI acceptance through psychological mechanisms. Two scenario-based experiments reveal a negative total effect of AI literacy: consumers with lower AI literacy show higher acceptance of AI technologies. Study 1 demonstrates that this effect is fully mediated by perceived mystery and awe, such that limited AI literacy increases mystification of AI behavior, evokes awe, and enhances acceptance. Study 2 further shows that AI anthropomorphism moderates this process. Higher anthropomorphism reduces perceived mystery among low-literacy consumers, thereby weakening the mystery–awe pathway. These findings challenge the assumption that increasing AI literacy necessarily promotes AI acceptance. The study contributes theoretically by uncovering affective mechanisms underlying AI acceptance and by extending Computers Are Social Actors (CASA) theory through identifying AI literacy as a boundary condition. Practically, the results provide insights for segment-specific AI interface design and responsible AI deployment.