VietTESOL International Convention 2026

This study investigates the pedagogical integration of AI-based feedback systems in high school writing instruction, focusing on their impact on learners’ writing performance and autonomy. Grounded in Sociocultural Theory, Constructivism, and Self-Regulated Learning, the research proposes the AI-Scaffolded Autonomous Writing (A-SAW) model as a framework for integrating ChatGPT-supported feedback into English writing classrooms. Using a mixed-methods, quasi-experimental design, the study involves Grade 11 students from rural public high schools who participate in AI-assisted writing activities over an extended instructional period. Quantitative data are collected through pre- and post-writing assessments and learner autonomy scales, while qualitative insights are obtained from interviews, reflective journals, and classroom observations. Preliminary findings indicate that AI-generated feedback promotes greater student engagement, revision awareness, and self-regulated learning strategies compared to traditional teacher feedback alone. The study also highlights the role of teachers in facilitating critical engagement with AI suggestions to prevent overreliance and encourage reflective learning. By offering an empirically grounded and pedagogically informed framework, this research contributes practical strategies for ethical and effective AI integration in secondary writing instruction and provides implications for teacher development, assessment design, and educational policy in AI-enhanced learning environments.
Key words: AI feedback, witing, writing instruction, writing performance, autonomy