The rapid emergence of ChatGPT has significantly reshaped research on academic writing within higher education, particularly in second language (L2) and EFL contexts. Despite the growing volume of studies, a systematic synthesis of how this research domain has evolved remains limited. This study aims to map the conceptual and social development of ChatGPT-related scholarship in academic writing through a bibliometric analysis of Scopus-indexed journal articles published between 2023 and early 2025. The relevant metadata, including citation information, bibliographic details, abstracts, and author keywords, were exported in RIS format and imported into VOSviewer for analysis. The analysis examined publication trends, dominant research themes, theoretical frameworks, methodological orientations, and patterns of scholarly collaboration. The findings reveal a clear developmental trajectory in the literature. Early studies primarily focused on academic integrity, plagiarism, authorship, and the detectability of AI-generated text. From 2024 onward, research increasingly emphasised pedagogical integration, highlighting ChatGPT’s potential to support linguistic accuracy, rhetorical organisation, and revision processes. More recent studies extend this focus to assessment practices, institutional governance, personalization, metacognition, and self-regulated learning, reflecting the growing conceptual maturity of AI-assisted academic writing research.
Binnendyk et al. (Tue,) studied this question.