Quantifying the digital divide: The relationship between AI dependency and academic integrity in EFL learners

As generative artificial intelligence (GenAI) becomes deeply embedded in higher education, concerns regarding student reliance on these tools have intensified. This study examined the predictive relationship between Academic Integrity (AIS) and Generative AI Dependency (GAID) among university EFL students (N = 117). Utilizing a quantitative correlational design, a simple linear regression revealed that academic integrity was not a significant predictor of AI dependency, F(1, 115) = 0.201, p = .655, accounting for a negligible 0.2% of the variance. Furthermore, one-way ANOVAs confirmed that both constructs remained stable across gender, age, major, and grade level (p > .05). Descriptive data on usage patterns provided a functional explanation for these null findings: students primarily utilize GenAI for idea generation (53.8%), linguistic correction (47.9%), and translation (41.0%). These results suggest a psychological decoupling where students categorize AI as a neutral cognitive scaffold rather than an ethical hazard. The study concludes that because dependency is driven by pragmatic utility rather than a lack of moral character, institutional responses must transition from traditional punitive integrity policies toward pedagogical frameworks that prioritize AI literacy and cognitive independence.