Integrasi Large Language Models Dalam Pembelajaran Analisis Teks Sastra Arab Di Perguruan Tinggi

Abstract

This study aims to analyze the impact of using Large Language Models (LLMs) on students’ ability to analyze Arabic literary texts. The research employed an experimental method with a pre-test and post-test design, involving an experimental class that used LLMs as analytical tools and a control class that followed conventional methods. Data were collected through analysis skill tests, students’ written products, questionnaires, and in-depth interviews. The findings indicate that the use of LLMs significantly enhances students’ ability to identify intrinsic elements, understand plot structures, interpret symbols and themes, and construct more coherent and systematic analytical arguments. In addition, LLMs facilitate the understanding of linguistic, cultural, and structural contexts within texts, and stimulate critical thinking through the evaluation of alternative interpretations and reflection on AI-generated answers. Students’ perceptions of LLM use were generally positive, particularly regarding ease of access to information, increased learning motivation, and accelerated comprehension of texts. Nevertheless, the study also identifies potential risks such as dependency, homogenization of analytical styles, and the limitations of AI in capturing cultural and aesthetic nuances of literary texts, indicating the continued need for instructor guidance and AI literacy. Overall, LLMs are proven to be effective pedagogical tools for improving the quality of Arabic literature learning, provided that their use is structured, ethical, and accompanied by academic supervision and reflective practice.