Automated Evaluation of ESL Learners’ English Writing Skills in English-Medium Instruction (EMI) through AI Writing Analytics

Abstract

This study analyses the use of artificial intelligence (AI)-driven automated writing evaluation (AWE) analytics and examines their role in enhancing the English writing skills of ESL learners in English-medium instruction (EMI) contexts. The review synthesises contemporary AWE systems, including those based on deep learning (DL), natural language processing (NLP), and generative AI approaches. It provides a detailed discussion of methods used for real-time feedback delivery, linguistic feature extraction, and the integration of AI-driven assessment with traditional teacher and peer feedback practices. In addition, the study critically evaluates empirical findings that highlight both the benefits and limitations of AI writing analytics in higher education EMI settings, including pedagogical, technical, and ethical challenges. Finally, the paper identifies future research directions and underscores the need for hybrid evaluation models that combine human oversight with automated systems to support technical accuracy, systematic assessment, and the development of higher-order writing skills in EMI contexts.