AI-Empowered English Speaking Teaching and Learning: A Scoping Review of Chinese Core Journal Studies (2022–2026)
TL;DR
By providing the first systematic mapping of this field, the review demonstrates how China's exam culture, in which high-stakes assessment pressure shapes AI integration in distinctive ways, interacts with AI+Education policy and the domestic LLM ecosystem to extend CALL and SLA scholarship beyond Western-centered settings.
Abstract
Despite rapid growth in research on generative AI in EFL speaking instruction, the thematic structure and evidence gaps of this body of research have not been systematically mapped. This scoping review follows PRISMA-ScR guidelines to map AI-empowered English speaking research in Chinese core journals indexed in CNKI from 2022 to March 2026. Two independent researchers screened the literature (Cohen’s κ = 0.85), yielding 33 eligible studies. Findings show that the field has grown substantially since 2024, shifting from tool-oriented discussions to systemic analyses across five themes: pedagogical restructuring, learner agency, dynamic assessment, teacher AI literacy, and ethical governance. Yet proficiency-stratified analysis, longitudinal evidence, assessment fairness, and comparative research on domestic Chinese and global LLMs remain underdeveloped. By providing the first systematic mapping of this field, the review demonstrates how China's exam culture, in which high-stakes assessment pressure shapes AI integration in distinctive ways, interacts with AI+Education policy and the domestic LLM ecosystem to extend CALL and SLA scholarship beyond Western-centered settings. Priority should be given to longitudinal and proficiency-stratified designs, integrated teaching-learning-assessment cycles, and localized LLM governance.