Generative AI Tools in Libyan Higher Education

Abstract

Generative artificial intelligence (Gen AI) tools are becoming increasingly common in students’ academic practices in higher education. However, evidence regarding students’ use, perceptions, and concerns about Gen AI tools remains limited. This study investigated the academic usage patterns, perceptions, and reported challenges associated with ChatGPT, Google Gemini, and Claude among undergraduate students at the Faculty of Information Technology, Alasmarya Islamic University (AIU), Zliten, Libya. This study employed a comparative cross-sectional quantitative analysis. Data were collected via an online questionnaire during the spring semester of 2025/2026 from 80 final-year female students enrolled in Computer Science and Internet Technology. Data were analysed using descriptive statistics, Cronbach’s alpha, and chi-square goodness-of-fit tests. The results showed that all participants used ChatGPT (100%), followed by Google Gemini (82.5%) and Claude (25.0%). Students reported using Gen AI tools for academic purposes, and the results also indicate challenges related to writing prompts, concerns about academic integrity and plagiarism, language and financial barriers, over-reliance on AI, and a lack of institutional guidance. The comparison results showed that ChatGPT was chosen most frequently for its accessibility and availability, ease of use, and usefulness for specific tasks. Google Gemini and Claude were chosen more frequently for output quality and reliability. Chi-square tests indicated that preferences differed significantly across the three tools on all four evaluation dimensions (p ≤ 0.001). The results suggest that Gen AI tools are already integrated into students' academic practices and locally provide evidence to support the development of institutional guidelines and the responsible use of Gen AI tools in Libyan higher education.