The Drivers of University Students' GenAI Dependency in Learning

Data were collected through Wenjuanxing.The final sample consisted of 1,384 valid responses after data cleaning.This study employed a structured questionnaire comprising five main sections. The first section collected demographic information including gender, grade, and major. The subsequent sections measured core constructs using established scales: a 6-item scale for AI dependency in learning(AIDL) was adapted from Ye et al. (2024); the 12-item AI literacy(AIL) scale from Wang et al. (2022) covering awareness, usage, evaluation, and ethics dimensions; along with 8-item perceived risk(PR) and 4-item perceived benefit(PB) scales regarding GenAI use in learning, both revised from Ivanov et al. (2024). All items used a five-point Likert scale format (1 = strongly disagree to 5 = strongly agree). To ensure response quality, two reverse-scored items were included in the AI literacy section. All scales demonstrated satisfactory reliability in the current research context.