Beyond Access: Socioeconomic, Academic and Disciplinary Determinants of Generative AI Use in Higher Education

Abstract

Although generative artificial intelligence (GenAI) is swiftly revolutionizing higher education, it prompted significant inquiries on adoption trends, equity and academic consequences. This research investigates the factors influencing the utilization of generative AI among university students, emphasizing demographic, academic, socioeconomic and technological variables. Data from a survey of 333 students were analyzed through cross-tabulations and chi-square tests to evaluate associations between the frequency of GenAI usage and variables such as gender, academic discipline, residence, parental education, household income, internet access and academic achievement. The findings demonstrate extensive adoption of GenAI, revealing notable disparities based on gender and academic specialty. Male students and individuals studying natural sciences indicated elevated levels of continuous usage. Socioeconomic characteristics, especially household income and parental education, significantly predicted both internet access and the frequency of generative AI usage, highlighting enduring digital disparities beyond mere access. GenAI usage was favorably connected with academic performance, suggesting it can boost academic success when correctly applied. Results show that discipline-specific and equity-oriented institutional regulations are needed for effective and accountable AI implementation in higher education.