Utilizing Generative Language Models (ChatGPT) to Enhance Writing Skills in Arabic Language Learning

Abstract

This study investigates the utilization of generative language models, specifically ChatGPT, as a pedagogical tool for improving Arabic writing skills. In the digital era, large language models (LLMs) provide new avenues for interaction, feedback, and creativity in language learning. This research explores how ChatGPT can facilitate scaffolding, idea generation, and linguistic correction in Arabic writing instruction. Using a qualitative descriptive design, data were gathered from classroom observations, student writings, and interviews with educators integrating ChatGPT into Arabic courses at Universitas Negeri Malang. Findings indicate that ChatGPT significantly enhances learners’ ability to organize ideas, expand vocabulary, and improve grammatical accuracy through interactive feedback and guided revision. However, the study also highlights potential drawbacks, including overreliance on AI output, reduced critical thinking, and ethical issues related to authorship and plagiarism. The research concludes by proposing a blended pedagogical model that integrates ChatGPT with teacher supervision and metacognitive reflection to ensure authentic, creative, and ethical Arabic writing practices.