Understanding Dual Pathways from AI Literacy to AI Reliance: A Two-Mediator Serial Mediation Model of Perceived Transparency and Trust

This paper examines how a users’ AI literacy shapes their reliance on AI communicative partners. Specifically, drawing on the Computers Are Social Actors (CASA) paradigm and the Heuristic-Systematic Model (HSM), the study investigates how AI literacy differentially influences two distinct interaction patterns: cautious reliance (critical engagement) and thoughtless reliance (passive acceptance). We contend that AI literacy influences these behaviors indirectly through a serial mediation pathway involving two key relational mechanisms: perceived transparency and AI trust. A cross-sectional survey was administered to 508 users of AI chatbots for health information, and the proposed serial mediation model was tested using structural equation modeling (SEM). Results indicated a significant suppression effect, revealing that AI literacy has competing direct and indirect effects on reliance. While AI literacy directly increased cautious reliance, it also indirectly decreased cautious reliance and increased thoughtless reliance via the serial pathway of enhanced perceived transparency and subsequent trust. This suggests that as users become more competent, the relational comfort and trust they build with their AI partner can paradoxically undermine their critical vigilance. Theoretical and practical implications are discussed regarding how human-machine communication processes can create complex and sometimes contradictory user behaviors, and how to foster more critically-minded AI engagement.