Abstract
INTRODUCTION: The rapid adoption of Generative AI (GenAI) in higher education raises concerns about psychological dependency. Grounded in the I-PACE model, this study investigates how lower need for cognition (NFC) is associated with problematic use via positive affect and Avoidance-Oriented GenAI Motivation, moderated by academic stress and AI literacy. METHODS: To test the hypothesized model, we employed a two-wave, time-lagged survey design with a sample of university students (N = 452). Data were analyzed using structural equation modeling (SEM) to assess the serial mediation effects and the moderated mediation dynamics. RESULTS: Analysis confirmed a serial mediation chain: lower NFC predicted stronger positive affect, increasing avoidance motivation and subsequent problematic use. This pathway was significantly amplified by high academic stress but attenuated by high AI literacy, which neutralized dependency risks. DISCUSSION: Findings extend the I-PACE model to instrumental technologies, identifying a specific "cognitive relief to avoidance" mechanism. The study highlights AI literacy as a vital protective resource, suggesting educators should prioritize fostering critical digital competencies over prohibitive policies to mitigate reliance risks.