Multi group analysis of demographic differences in higher education students ChatGPT use behaviour within a modified UTAUT2 model

Although extensive research has examined AI integration in education and the UTAUT2 model, few studies have explored ChatGPT adoption with demographic variables in Nigeria and other African contexts. Prior studies have analyzed the impacts UTAUT variables have on students’ behavioral intention and use of ChatGPT, yet little is known about how these relationships differ by demographic profiles. This study addresses that gap by investigating the influence of factors such as age, sex, and programme of study on these associations among Nigerian higher education students. A cross-sectional correlational design was used and data were collected from 8,496 students across various tertiary institutions in Nigeria, using a structured questionnaire. Data were gathered electronically between October 13, 2023, and February 14, 2024. Multigroup analysis within the framework of Partial Least Squares Structural Equation Modelling (PLS- SEM) was performed. Both measurement and structural model metrics are reported for all subgroups. Findings indicate that performance expectancy is associated with higher intention to use ChatGPT, particularly among males, older students, and those with advanced qualifications, while its relationship with actual use varies across age groups. Effort expectancy and social influence showed differing associations with intention and use depending on demographic characteristics. Facilitating conditions were generally linked to greater actual use, and both hedonic motivation and habit were consistently associated with intention and behavior. These results provide empirical evidence of demographic variations in ChatGPT adoption among Nigerian students and suggest that interventions promoting AI tools should consider these differences.