Adoption of Generative AI in Portuguese Higher Education: An Empirical Study of Perceptions and Academic Practices

Abstract

Generative artificial intelligence (GenAI) is diffusing rapidly in higher education, yet empirical evidence from Portugal remains limited. We conducted an online questionnaire (n = 77) targeting higher-education faculty and researchers in Portugal and analyzed the data using descriptive statistics and non-parametric tests (Mann–Whitney, Kruskal–Wallis). Participants most frequently reported using QuillBot, followed by Bing Chat and ChatPDF, with a sizable minority indicating the use of additional, unlisted GenAI applications and reported use cases clustered around idea generation, drafting, summarization, proofreading, and coding, while trust and appropriateness varied by task type. Inferential tests suggested differences by gender and by discipline for selected items (e.g., p-value=0.008, p-value=0.039, p-value=0.023); given the small and imbalanced subgroups, these findings should be interpreted cautiously and supplemented with effect sizes and confidence intervals. No statistically significant differences by age or gender were observed for the ethical-concerns items. We discuss implications for training, assessment, and policy design, and outline pragmatic recommendations for responsible GenAI integration in Portuguese higher education.