Toward Responsible Integration of GenAI in Arabic Teaching: Insights from Thematic Analysis
Mozah Hamad Al Kaabi, Mohamed Bin Zayed University for Humanities (United Arab Emirates)
Asma Saeed Almaamari, Mohamed Bin Zayed University for Humanities (United Arab Emirates)
Abstract
Artificial Intelligence (AI) systems in Arabic remain markedly less advanced than those designed for more standardized languages such as English. The challenges stem largely from the linguistic complexity of Arabic grammar, the diversity of regional dialects, and the scarcity of high-quality datasets. Against this backdrop, the present study examines the role of Generative AI (GenAI) as a virtual teaching assistant in Arabic language instruction at Mohamed Bin Zayed University for Humanities. Employing a qualitative research design, we conducted semi-structured interviews with 15 instructors and analyzed the data thematically. Findings indicate that instructors utilize GenAI for instructional material development, student assessment, and the design of personalized learning pathways. Nevertheless, persistent difficulties were reported regarding the reliability of GenAI in handling dialectal variation, maintaining cultural authenticity, and supporting valid assessment practices. Students, in turn, employ GenAI for writing practice, vocabulary acquisition, and grammar support, yet express concerns over issues of appropriate use and academic integrity. The analysis further highlights gaps in teacher training, assessment frameworks, and institutional policies governing AI integration. The study concludes by recommending the formulation of Arabic-specific AI guidelines, the development of innovative assessment strategies, and the establishment of clear institutional protocols to ensure responsible and pedagogically sound applications of GenAI in Arabic language education.
Keywords: Arabic Language Teaching; Generative AI; Virtual Teaching Assistant; Assessment Strategies; Teacher Training; Cultural Authenticity