Mapping thematic evolution and pedagogical strategies of Generative AI in EFL writing

Abstract

The mass adoption of Generative AI has brought a massive landscape shift in academic writing for EFL students. This technology acts as a powerful cognitive scaffold for students on finishing their academic writing tasks such as brainstorming ideas, checking grammar accuracy, paraphrasing a sentence and as sparring partner. Nevertheless, the tools also act as double-edged sword because its comprehensive aid that triggers the scaffolding paradox, cognitive offloading, ethical issue or academic integrity. However, these perspectives remain distributed across several research domains, making it harder to understand how the field is developing as an integrated research area. The study employs a hybrid study design using bibliometric informed scoping review to examine the thematic structure and the development of research on generative AI in EFL academic writing. The research aims to quantitatively map the thematic evolution of AI in EFL writing globally from 430 Scopus-indexed articles from January 2023 to June 2026, and also qualitatively synthesize the condition when AI act as substitute or act as cognitive behavior mechanism of overreliance on using Generative AI, and aims to offer a strategy or framework about practical AI literacy in EFL academic writing. Rather than treating Generative AI as inherently beneficial or harmful, it examines the emerging tension between AI as cognitive assistance and AI as a potential substitute for EFL learners. The study mainly proposes an evidence-informed conceptual framework that conceptualizes effective Generative AI assisted writing as a process in which this tool remains subordinate to learner critical thinking and judgment.