Artificial Intelligence in English-Speaking Education: A WoS-Based Bibliometric Analysis of Technology, Pedagogy, and Ethics (2021-2025)

Abstract

This study provides a bibliometric overview of artificial intelligence (AI)-assisted English-speaking education based on 248 WoSindexed articles published between 2021 and 2025 (2,046 citations; Hindex = 24). Using co-citation, co-occurrence, and temporal analyses, the study identifies major research patterns. The findings grounded and shift from an early emphasis on speech technologies to learner-centered, pedagogically grounded, and affect-sensitive approaches. The analysis identifies a critical gap, as ethical and governance concerns remain insufficiently addressed. Ethics-related terms do not appear among highfrequency or high-centrality keywords, despite their growing relevance in AI-enhanced learning. By synthesizing structural and temporal patterns, the study clarifies interactions among technological, pedagogical, and learner-related dimensions, while highlighting the under-theorization of ethics and governance. It also reveals a fragmented use of theoretical models. Frameworks such as the Technology Acceptance Model (TAM), Social Cognitive Theory (SCT), and SelfDetermination Theory (SDT) are typically applied in isolation rather than being integratively linked to instructional design. Beyond descriptive mapping, this study advances an integrative perspective linking technology, pedagogy, and learner experience, with implications for teachers, researchers, and policymakers. The findings align with UNESCO’s SDG 4, highlighting the need for inclusive and equitable AI-supported speaking education.