Turkish EFL teachers’ perspectives on AI-generated feedback: Negotiating trust, control, and pedagogical adaptation in writing instruction

Abstract

This qualitative study explored how EFL teachers in Turkish higher education perceive, adapt to, and manage AI-generated feedback in writing instruction. Drawing on Teacher Cognition Theory and the Technology Acceptance Model, the research examined how thirty-six university instructors negotiated issues of pedagogical value, professional control, and trust in AI-mediated feedback. Data were gathered through an open-ended questionnaire during the second academic term of 2025 and were analysed thematically. Three major themes emerged: teachers’ perceptions of the pedagogical value and reliability of AI feedback, their adaptation of feedback practices and instructional decisions, and the professional challenges arising from the integration of AI. The findings indicate that teachers regarded AI feedback as a valuable supplement for improving linguistic precision and efficiency, while maintaining scepticism about its interpretive reliability and ethical adequacy. Their adaptations included filtering AI-generated suggestions, refining assessment rubrics, and promoting reflective discussion to preserve pedagogical judgement. Concerns centred on student dependence, the erosion of teacher authority, and the need for structured professional preparation in the critical use of AI. The study advances the theoretical understanding of teacher cognition and technology acceptance in the context of intelligent feedback systems, offering practical implications for designing teacher-training programmes and institutional policies that sustain a reflective, responsible, and contextually grounded integration of AI in EFL writing pedagogy.