Perceptions of EFL Instructors on ChatGPT as a Co-Agent: An Exploratory Study of Classroom-Based Writing Activities
JOURNAL OF LANGUAGE TEACHING AND LEARNING, cilt.16, sa.2, ss.52-66, 2026 (ESCI)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 16 Sayı: 2
- Basım Tarihi: 2026
- Doi Numarası: 10.66887/jltl.v16i3.1072
- Dergi Adı: JOURNAL OF LANGUAGE TEACHING AND LEARNING
- Derginin Tarandığı İndeksler: Emerging Sources Citation Index (ESCI), MLA - Modern Language Association Database, MLA International Bibliography, EBSCO Communication Source, Communication Source (EBSCO)
- Sayfa Sayıları: ss.52-66
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Gazi Üniversitesi Adresli: Evet
Özet
Generative AI, especially ChatGPT, is increasingly used for writing feedback in higher education EFL classrooms. Although previous studies have noted its immediacy and comprehensiveness, concerns about accuracy and alignment have also been raised. Existing studies often emphasize technological capability over how AI feedback is managed in instruction. This study examined how ChatGPT can be positioned as a feedback agent in EFL writing and how co-agency is negotiated within the network of teachers, students, and ChatGPT. Thematic analysis was conducted on written interviews of 11 EFL teachers. The findings revealed that ChatGPT was not positioned as an author in its own right. Instead, it was described as a conditional co-agent. ChatGPT is most useful when integrated into teacher-led practice and aligned with task criteria. Because co-agency develops through classroom routines, whether ChatGPT supports writing or creates friction depends on how instructors position and regulate its use. (c) Association of Applied Linguistics. All rights reserved