AI-Enabled Assessment in Language Learning: Design, Implementation, and Effectiveness

Artificial intelligence (AI) is increasingly integrated into language education, yet its application in assessment remains comparatively underexplored. While prior research has largely focused on AI-supported instruction, less systematic attention has been given to how AI-enabled assessment systems are designed, implemented, and evaluated for effectiveness. This systematic review synthesizes evidence from 50 high-quality studies selected through a PRISMA-informed methodology from an initial pool of 961 records. The review examines three core dimensions: (a) the design of AI-enabled language assessment tools, (b) their implementation and institutional integration, and (c) their effectiveness in supporting learning outcomes and assessment quality. Findings indicate that AI-enabled assessments—particularly automated feedback, scoring, and adaptive systems—offer benefits in efficiency, learner engagement, and formative learning. However, challenges related to validity, fairness, transparency, and ethical governance persist. The review identifies critical research gaps in longitudinal studies, school-level contexts, multimodal assessment, and policy frameworks. Implications for educators, institutions, and policymakers are discussed, emphasizing the need for responsible, pedagogically grounded, and ethically informed adoption of AI-enabled assessment in language learning.