The Pedagogical Value of AI-Based Writing Feedback: Pre-University Learners’ Perceptions and Writing Improvement

The integration of artificial intelligence (AI) into English language instruction has transformed pedagogical practices, particularly in writing instruction and feedback provision. Despite increasing scholarly attention to AI-assisted writing support, limited research has explored the pedagogical value of AI-generated feedback among pre-university English as Foreign Language (EFL) learners, especially in developing educational contexts. This mixed-methods study investigates pre-university learners’ perceptions of AI-based writing feedback and examines its influence on writing improvement. The study was conducted among foundation-level A2 English learners in a pre-university setting, with a population of 200 students and a purposive sample of 31 participants aged 17–19 years. A hybrid feedback model combining AI-generated feedback and teacher mediation was implemented. Data were collected through a Likert-scale questionnaire, semi-structured interviews, and writing performance evidence obtained from pre-feedback and revised writing drafts. Quantitative data were analysed using descriptive statistics and reliability testing, while qualitative data underwent thematic analysis. Writing performance was evaluated using a rubric focusing on grammar, vocabulary, organization, coherence, and mechanics. Findings indicate that learners perceived AI-based feedback positively, particularly regarding grammar correction, vocabulary enhancement, and revision support. Students also valued teacher mediation for contextual clarification and emotional reassurance. Writing evidence demonstrated measurable improvements in organization, mechanics, and coherence following AI-supported revisions. The study highlights the pedagogical potential of hybrid AI–teacher feedback in pre-university EFL writing classrooms and offers implications for integrating AI tools into language instruction while maintaining teacher oversight.