Artificial intelligence in language assessment research: a scoping review

Artificial Intelligence (AI) transforms language assessment through applications such as automated scoring, item generation, and test administration. As research in this area expands rapidly, language assessment specialists have struggled to keep pace. While synthesis studies have appeared in related fields such as language education and educational measurement, a focused, up-to-date review of empirical research within language assessment remains lacking. This study addresses that gap by reporting on a scoping review of 212 studies published between 2019 and 2024. The review explores how AI is used in language assessment research by examining publication characteristics, research themes, targeted language skills and knowledge areas, types of AI tools and non-AI instruments used, data analysis methods, AI-related implications, and construct reconceptualisation. Key findings include a dominant focus on automated scoring of speaking and writing and widespread reliance on off-the-shelf AI tools. Notable gaps were also identified, such as limited research on AI for test validation, assessment of receptive skills and sign languages, and the lack of guidance for stakeholders on appropriate AI use. The review concludes by outlining directions for future research in language assessment in an AI-driven context.