When Usefulness Fuels Fear: The Paradox of Generative AI Dependence and the Mitigating Role of AI Literacy

Artificial intelligence (AI) literacy is essential for understanding and benefiting from AI, yet little is known about how users form misconceptions of AI and how AI literacy (AIL) mitigates them. Adopting media dependency theory, this study examines factors influencing dependence on generative AI (DGAI) and its relationship with fear of AI (FAI) across different AIL levels. The in-depth interviews reveal three stages of developing DGAI: tool acceptance, habit formation, and psychological dependence. This dependency raises concerns about privacy, skill degradation, and job displacement. Survey results show that perceived risk directly increases FAI, while perceived usefulness and anthropomorphism indirectly affect FAI through DGAI, with self-efficacy reducing DGAI. Moderation analysis finds that AIL buffers the fear-inducing effects of usefulness and anthropomorphism, especially at low AIL levels. This study emphasizes the dynamic relationship between AIL, users' dependency, and attitudes toward generative AI, emphasizing AIL's role in reducing misconceptions and fostering sustainable AI development.