Two years of innovation: A systematic review of empirical generative AI research in language learning and teaching

This systematic review examines the evolution of empirical research on generative AI in language learning and teaching from 2023 to 2024. Following PRISMA guidelines, we analyzed 144 peer-reviewed articles from Web of Science, Scopus, and ERIC databases to explore the field's progression, including identifying new developments, shifts in research priorities, and emerging themes that have arisen compared to the first year. The findings reveal an exponential growth in publications, with a significant shift from primarily exploratory inquiries to more systematic and empirically driven investigations. The analysis identified six main research foci: Perceptions and attitudes, psychological and cognitive aspects, teaching and learning strategies, language skills development, writing and feedback, and implementation and integration. While higher education (86.7 %) and English as Foreign Language contexts (86.1 %) dominated the research landscape, there was notable geographical diversity, with strong representation from East Asia and emerging contributions from the Middle East. Mixed-methods approaches (38.9 %) were prevalent, with increasing incorporation of AI-generated content, multimedia recordings, and digital interaction data from 2023 to 2024. Writing emerged as the primary focus (42.4 %) while speaking, listening, and reading skills received comparatively less attention. This review highlights critical gaps, including limited research in K-12 settings, insufficient longitudinal studies, and the need for more diverse language representation beyond English. These findings suggest the field is maturing but requires broader investigation across educational levels, language domains, and geographical contexts to fully understand the impact of generative AI in language education.