Guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework, this review synthesizes 24 empirical studies between 2015 and 2025 on the role of artificial intelligence (AI) in shaping English learners’ willingness to communicate (WTC). The findings show that most studies were conducted between 2024 and 2025, and were heavily concentrated in East Asia, particularly Mainland China, likely reflecting supportive policies, technological infrastructure, and a rapid response to recent advances in educational AI. Methodologically, most studies adopted longitudinal and experimental designs with standard class-sized samples, offering controlled classroom-based insights. Across different AI types, AI-supported learning environments were found to support WTC primarily through affective, cognitive, and social mechanisms. Nevertheless, the review identifies persistent conceptual and pedagogical challenges, including heavy reliance on self-report measures, limited attention to interaction quality and contextual specificity, and insufficient consideration of ethical issues, which collectively constrain the interpretation of AI-supported WTC development. Accordingly, the review concludes that future research should move beyond tool-centered evaluations and adopt more theoretically grounded, measurement-sensitive, and context-responsive approaches, situated within teacher-guided pedagogical designs, to better understand how AI can sustainably support learners’ WTC.
Cao et al. (Wed,) studied this question.