Generative AI Overuse, Autonomy, and Mental Well-Being of Students at a Selected University in Kazakhstan

Abstract

Generative artificial intelligence is now part of everyday university learning. It offers speed, access, and support. Yet it also raises a more difficult question. When does useful assistance turn into overreliance, with consequences for students’ autonomy and well-being? Research on Gen AI in higher education has expanded rapidly, but most studies focus on patterns of use, attitudes, or academic integrity. Much less is known about Gen AI overuse and its relation to dispositional autonomy and mental well-being, especially in under-examined contexts such as Kazakhstan. This study addressed that gap. It examined levels of Gen AI overuse, dispositional autonomy, and mental well-being among university students in Kazakhstan and investigated their relationships. An explanatory sequential mixed methods design was employed. The quantitative phase involved a survey of 409 students. The qualitative phase included 13 semi-structured interviews to explain and deepen the quantitative results. The findings showed that Gen AI overuse was below the midpoint of the scale on average, while dispositional autonomy and mental well-being were above it. Gen AI overuse was negatively associated with both autonomy and well-being. Dispositional autonomy also partially mediated the relationship between overuse and well-being. Demographic and academic variables added little explanatory value beyond the psychological factors in predicting overuse. The interviews added depth to these findings. They showed that overreliance often developed through convenience, workload pressure, and academic demands. They also revealed reduced ownership of learning, weaker independent thinking, and emotional strain when reliance became excessive. Overall, the study shows that the central issue is not Gen AI use itself, but the point at which support begins to replace independent thought. The findings call for autonomy-supportive, well-being-oriented, and pedagogically responsible approaches to Gen AI in universities.