AI-Based Feedback in Academic Writing: Insights from EFL Undergraduate Students at Saudi Higher Education

Abstract

With the rapid adoption of artificial intelligence technologies in higher education, understanding students' attitudes toward these tools is essential for shaping effective academic practices. By grounding the investigation of AI-based feedback within socio-cognitive and self-regulated learning theories, this study examined undergraduate students' perceptions of artificial intelligence (AI) tools such as Grammarly, ChatGPT, and MS Copilot as feedback tools for their academic writing. The study adopted a quantitative, cross-sectional survey design to capture students' perceptions. The sample consisted of 100 undergraduate students in the College of Medicine at Taif University. Data were collected using an adapted structured questionnaire employing a five-point Likert scale spanning from 'strongly disagree' (1) to 'strongly agree' (5). Data were analyzed using SPSS (v. 26) to ensure the validity and reliability of the questionnaire, and to perform descriptive statistical tests. The findings revealed that students held positive perspectives of AI as a feedback tool on their writing performance, learning and reflection, and affective impact. Furthermore, students reflected a high level of trust in the quality and credibility of AI-generated feedback for supporting academic writing tasks. The study recommended integrating AI literacy into writing curricula, establishing institutional guidelines for responsible AI use, and encouraging balanced reliance on AI tools with the teachers' guidance.