Students' Perceptions and Usage Patterns of ChatGPT as an Automated Writing Evaluation (AWE) Tool

Abstract

The emergence of Generative AI (GAI) has significantly influenced writing evaluation, with recent literature primarily focusing on comparing GAI-generated feedback or scores with those provided by humans. However, fewer studies have explored students’ perspectives, which play a crucial role in the effectiveness of feedback. To address this gap, this study investigates how EFL students interact with ChatGPT as an Automated Writing Evaluation (AWE) tool and their perceptions of its usefulness in improving academic writing. The findings reveal that students primarily used ChatGPT during the revision stage for grammar and vocabulary enhancement, but its use for higher-order writing tasks, such as argumentation and content development, was limited. Moreover, students generally viewed ChatGPT as trustworthy, useful, and easy to use, though there were concerns about information security and ethical implications. There were also moderate anxiety about using the tool, highlighting the need for additional training and reassurance. Despite these challenges, students expressed a willingness to continue using ChatGPT, suggesting its potential as a supplementary tool for academic writing instruction. This study underscores the need to address user concerns and refine GAI tools to better support advanced writing tasks and foster ethical, confident use in educational contexts.