This study investigates how university students in Poland, Turkey, the United Kingdom, and the United States engage with generative artificial intelligence (GenAI) tools in their academic work, and how this usage influences their perceptions in two key areas: (1) performance expectancy, defined as the degree to which students believe that using GenAI will help them improve their academic performance, and (2) sustainable education, reflecting beliefs about GenAI’s potential to foster equitable, inclusive, and future-oriented learning environments. Drawing on survey data from 578 students, the study applied partial least squares structural equation modeling (PLS-SEM) using SmartPLS 4 to evaluate both the measurement and structural models. The findings indicate that more frequent use of GenAI tools is positively associated with students’ perceptions of their academic utility and their contribution to sustainability-oriented learning, particularly in Poland, Turkey, and the United States. However, no such associations were found in the United Kingdom, suggesting some contextual variation. Overall, limited cross-national differences were observed, pointing to a convergence in how students across diverse cultural settings experience and evaluate GenAI in education. Given the novelty of GenAI adoption in an academic context, further research is warranted to replicate and extend these findings across broader disciplinary and demographic samples.