Abstract

This systematic review examines the development and application of automated writing evaluation (AWE) in language writing instruction. AWE systems are increasingly used to reduce teachers’ workloads and provide learners with more frequent feedback. However, systematic syntheses of high-quality empirical research in this field remain limited. To address this gap, the review analyzes empirical studies published over the past decade, focusing on research methodologies, participants’ proficiency levels, and the benefits and challenges of using AWE. The findings indicate that AWE can improve overall writing performance, linguistic accuracy, lexical diversity, and syntactic complexity, particularly across multiple drafts. Nevertheless, AWE primarily addresses lower-order linguistic features and is less effective in providing feedback on content, organization, meaning, and communicative intent. The review therefore emphasizes that AWE should complement rather than replace teacher feedback. Future research should adopt more rigorous methodologies and examine genre-related differences, long-term effects, cognitive and affective engagement, effective pedagogical integration, and the development of advanced AWE tools capable of providing higher-order feedback.

Keywords: Automated feedback, automated writing evaluation, language writing, systematic review, writing performance.