Abstract
Social-emotional development plays a key role in second language acquisition (SLA). With recent advancements in artificial intelligence (AI), there is increasing interest in how AI technologies can enhance social-emotional factors in SLA. This study aims to systematically synthesize empirical research on the role of AI in shaping social-emotional experiences in SLA, identify prevailing trends, summarize patterns in reported outcomes, examine proposed mechanisms, and highlight directions for future research. Following PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines, this scoping review encompasses 66 empirical studies published between 2022 and April 2025. All studies were independently reviewed by two researchers and appraised using the Mixed Methods Appraisal Tool (MMAT). Results: (1) a marked increase in scholarly attention to AI-supported social-emotional development in SLA; (2) AI’s reported potential in enhancing motivation and reducing anxiety, but evidence is limited by positivity bias, methodological inconsistencies, and concentration in East Asian contexts; and (3) five reported mechanisms through which AI may support social-emotional growth in SLA: dialogic interaction and perceived social presence; perceived competence, agency, and task value; emotion regulation and timely affective support; immersion, flow, and sustained engagement; and reduced evaluative pressure and psychological safety. At the same time, the review identifies insufficient critical attention to potentially negative or constraining aspects of AI-supported language learning, including learner over-reliance and reduced autonomy. This review offers a balanced synthesis of the current evidence and highlights implications for future AI–SLA research, pedagogy, and design.