This study examines the role of artificial intelligence (AI) in EFL writing feedback, focusing on three dimensions: lower-order writing skills (grammar, vocabulary, and mechanics), higher-order writing skills (idea flow, coherence, and argumentation), and learner interaction with AI-generated feedback during revision. A qualitative thematic review of 45 peer-reviewed empirical studies published from 2023 to 2025 was conducted, following Braun and Clarke's (2006) six-phase thematic analysis framework with strict inclusion criteria. Thematic coding identified four overlapping themes: linguistic accuracy development, discourse-level writing enhancement, learner engagement and revision habits, and factors influencing trust, acceptance, and resistance toward AI feedback. The findings reveal that AI feedback effectively supports lower-order writing accuracy through immediate and iterative correction, while generative AI tools are increasingly contributing to higher-order writing development through discourse-level guidance. However, learner engagement with AI feedback remains selective, shaped by digital literacy, perceived relevance, trust, authorship, and control concerns. This study underscores the importance of feedback literacy instruction and advocates for a balanced integration of AI and human feedback to maximize pedagogical effectiveness in EFL writing instruction.
Md Fahid Hossain (2025) studied this question.