<![CDATA[Background: Artificial Intelligence (AI), particularly ChatGPT, has created new opportunities in language learning by providing automated written corrective feedback. However, studies examining students’ experiences in using ChatGPT as corrective feedback in Arabic writing skills instruction remain limited, particularly within the context of higher education in Indonesia. Aims: This study explores students’ experiences using ChatGPT as corrective feedback in Arabic writing instruction, identifying perceived benefits, challenges, and educational significance. Methods: This study employed a qualitative approach using a descriptive phenomenological research design. The participants were students of the Arabic Language Education Program, FITK UNSIQ Wonosobo, selected through purposive sampling. Data were collected through in-depth interviews and documentation, including students’ written assignments and screenshots of their interactions with ChatGPT. Data were analyzed using Colaizzi’s phenomenological analysis method. To ensure the trustworthiness of the findings, source triangulation, member checking, and peer debriefing were employed. Results: The findings revealed five major themes: (1) ChatGPT provides corrective feedback that is easy to understand through clear explanations; (2) it enhances students’ revision skills and learner autonomy; (3) it supports idea development and improves the quality of students’ writing; (4) it encourages students to verify the feedback due to concerns regarding its accuracy; and (5) it improves task completion efficiency by providing rapid feedback. Overall, the participants perceived ChatGPT as an effective learning companion; however, they emphasized that it cannot replace the role of lecturers in the teaching and learning process. Implications: The findings suggest ChatGPT has considerable potential as a corrective feedback tool to support Mahārah al-Kitābah instruction, particularly in developing revision skills and learner autonomy.]]>
