Artificial intelligence in higher education: Modelling the antecedents of artificial intelligence usage and effects on 21st century employability skills among postgraduate students in Ghana

The utilisation of Artificial intelligence (AI) in higher education has been a contemporary issue attracting varied perspectives. However, the nuances among the antecedents of AI usage among postgraduate students and the acquisition of 21st-century employability skills are under-researched. This study sought to fill the chasm by adopting a concurrent integrated mixed method to collect data using a questionnaire with closed-ended and open-ended items from 294 postgraduate students in Ghana. Descriptive statistics and partial least squares structural equation modelling (PLS-SEM) techniques were employed for the quantitative data analyses while thematic pattern matching was used to analyse the qualitative data gathered. Results from the descriptive analysis revealed that the most prevalent AI platforms used by postgraduate students were ChartGPT and Quillbot. The two most important uses of such AI tools were just used for learning in general; and searching for literature. The PLS-SEM paths modelling showed that antecedents of postgraduate students' AI usage were benefits, challenges and platform availability. Finally, both AI usage and AI challenges were significant predictors of postgraduate students’ acquisition of 21stcentury employability skills. However, the relationship between AI challenges and the acquisition of 21st-century employability skills was fully mediated by AI usage. Based on the results, recommendations were made for policy and practice of AI usage in higher education.