Artificial Intelligence in Language Learning: A Twenty-Year Scoping Review of Applications, Research Methods, and Outcomes

This scoping review examines the integration of artificial intelligence (AI) in language learning research over the past two decades (2005-2024). A total of 272 empirical studies were analyzed to identify emerging trends in AI technologies, research methodologies, and reported language learning outcomes. The findings reveal a significant shift from rule-based systems to advanced AI applications, with a marked predominance of chatbots (44.1%) and commercial tools (71%) in recent research. Methodologically, there has been a strong emphasis on productive skills, particularly writing (93 studies) and speaking (61 studies), alongside growing attention to affective learning outcomes (98 studies). While mixed methods approach (53.5%) dominated the research context, notable limitations include the overrepresentation of university students (67.6%), formal learning contexts (76.5%), and English as the target language (93.0%). The overwhelmingly positive outcomes reported (79.5%) raise important questions about publication bias and methodological rigor. This review identifies critical gaps in current research, including the limited attention to receptive skills, non-English languages, and informal learning environments. The study proposes future research directions that address these gaps through methodological diversification, expanded contexts, and more comprehensive effectiveness measures to better understand how AI can effectively support language acquisition across diverse learning environments.