Document Type : Original Article
Authors
1 Assistant Professor, Department of Arabic Language and Literature, Faculty of Literature and Humanities, Zabol University, Zabol, Iran.
2 Master's degree student, Department of Arabic Language and Literature, University of Zabol, Zabol
Abstract
Amid the accelerating digital transformation in higher education, there is a growing need to evaluate the efficacy of generative artificial intelligence (AI) models in enhancing the quality of formative feedback. This study aimed to explore the impact of using DeepSeek, a Chinese open-source large language model, for integrating AI into the process of assessing Arabic composition (Insha’). Employing a qualitative-analytical methodology, the study focused on a purposive sample of six faculty members and 32 undergraduate students in the Arabic Language Department at Zabol University. The results indicated that the program contributed to improving the temporal efficiency of assessment and enriched the quality of feedback by providing comprehensive linguistic analyses (grammatical, morphological, lexical), thereby enhancing iterative learning among students. The study also revealed significant challenges, most notably dependence on stable digital infrastructure, the need for precise prompt engineering skills, and concerns related to cultural appropriateness and academic privacy. The study concludes that the optimal deployment of DeepSeek should occur within an integrative pedagogical framework, where it functions as an intelligent assistant to bolster the student's capacity for critical thinking and self-review, rather than serving as a substitute for it. This research offers a unique applied contribution to the field of Arabic writing instruction through a dual-level interactive analysis (instructor-student) within an authentic university context.
Keywords
)