<![CDATA[The limited academic writing skills of undergraduate students often hinder their ability to produce coherent, critical, and argumentative scholarly work. The lack of adaptive interventions in writing instruction further exacerbates this challenge, particularly in the digital era that demands the integration of technology. This study aims to analyze the effectiveness of ChatGPT in enhancing students’ academic writing quality, explore students’ perceptions of its use, and examine the emerging ethical implications. Employing a mixed-methods design, the study involved 114 undergraduate students who participated in AI-assisted writing training. Data were collected through writing tests (pre-test and post-test), closed-ended questionnaires, and open-ended interviews. Research instruments included an academic writing rubric and a Likert-scale perception survey. Data analysis was conducted using paired t-tests for quantitative data and thematic analysis for qualitative data. The findings revealed a significant improvement of 29.2% in students’ academic writing performance. Furthermore, 92% of respondents perceived ChatGPT as effective in supporting idea development and paragraph organization. However, concerns were also raised, with 65% reporting potential dependency and 48% struggling to distinguish between AI generated content and their own work. These results indicate that ChatGPT has the potential to serve as an innovative learning tool in academic writing, yet its application requires appropriate pedagogical strategies and regulatory frameworks to minimize ethical risks in higher education.]]>
Copyrights © 2025
