Mapping the Evolution of AI Feedback in ESL/EFL Writing: A Chronological Bibliometric Analysis

Abstract

Despite the rapid integration of artificial intelligence (AI) feedback into ESL/EFL writing pedagogy, research providing a comprehensive overview of how the field has evolved, diversified, and matured remains limited. This study examines the development of AI feedback research in ESL/EFL writing from 2000 to 2025 through a chronological bibliometric analysis. Records were retrieved from Scopus and the Web of Science Core Collection and screened through a PRISMA-style process involving identification, deduplication, title and abstract screening, and full-text eligibility assessment. After screening, 282 empirical studies were included in the final dataset. Descriptive statistics and VOSviewer-based visualizations were used to analyze publication growth, major journals, countries/regions, institutions, highly cited studies, productive authors, keyword patterns, AI tools, feedback-focus categories, and methodological features. The results show that the field remained limited for many years but expanded sharply in 2024 and 2025. Research output was concentrated in a small group of journals and showed a strong presence of Asian EFL contexts, especially China and Hong Kong. The findings also indicate that the field is shaped by both earlier automated writing evaluation traditions and newer generative AI research. ChatGPT was the most frequently reported tool, while grammar, organization, vocabulary, and content were the most common feedback focuses. Methodologically, mixed-method research and quasi-experimental designs were especially visible. The study provides a structured overview of the field and points to the need for more transparent reporting of AI tools, feedback types, revision procedures, and higher-level writing features such as cohesion, coherence, argumentation, and development.