The Role of Generative AI in Enhancing or Undermining Autonomous Learning in Higher Education

Rizwan Shoukat, Government College University, Faisalabad
Khalida Parveen*, Qujing Normal University
Ashraf Iqbal, Government College University, Faisalabad
Muhammad Shafiq, Qujing Normal University
Huang Qibing, Soochow University

https://doi.org/10.9743/JEO.2026.23.1.12

Abstract

This paper analyses the effect of generative AI on the improvement or deterioration of autonomous learning in higher education. Given that AI tools, such as ChatGPT and DALL-E, are increasingly being incorporated into academic practices, it is vital to understand their effects on student autonomy. The application of SEM and CFA can help to better understand the connections between AI use, student attitude, and dependency. This study had a sample size of 100 higher education students with demographic variability (50% men and 50% women; between the ages of 19 and 32; 48% undergraduate, 32% master’s, and 20% PhD students). Results show that intensive AI use develops students’ autonomy perceptions (path coefficient = 0.6–0.7, p < 0.05) but negatively relates to autonomy if dependency on AI is overused (path coefficient = −0.5, p < 0.05). Correlation analysis showed a positive relationship between AI usage and attitudes (r = 0.45), but a moderate correlation with dependency (r = 0.40), which brings into question overdependence. The findings, therefore, present a multifaceted role for generative AI in higher education that calls for a balanced usage and guidance at the institutional level.

Keywords: generative AI, autonomous learning, student perceptions, educational technology, AI literacy

*Corresponding Author: nicepak81@hotmail.com