Development of a Diagnostic Tool for Measuring College Students' Dependence on Generative AI: A Case Study of College K

This study was conducted to develop and validate a diagnostic tool to multi-dimensionally measure college students' dependence on the use of generative AI in a generative AI environment. Following a literature review and expert consultation, 14 preliminary items were constructed, and exploratory and confirmatory factor analyses were performed based on responses from 557 undergraduate students at University K. As a result of the study, a total of 10 items with a two-factor structure: cognitive dependence (4 items) and emotional and behavioral dependence (6 items) were finalized. The KMO value of the developed tool was found to be excellent, and the model fit was very good, thereby securing structural validity. Unlike previous scales centered on a single factor, this study has academic significance in that it reflects the unique characteristics of generative AI dependence, namely the “outsourcing of thinking”. It can serve as basic data for understanding the current status in university education settings and for designing customized educational programs.