Exploring AI-Driven Written English Assessment: Toward Improved Assessment Quality and Learner Outcomes

The use of AI-powered tools for assessing written English is revolutionizing conventional evaluation methods by improving accuracy, consistency, and teaching efficiency. However, there is limited research on how AI assessment tools affect the validity and fairness of language evaluation. This review synthesizes current research on the technological foundations, pedagogical impact, and ethical considerations of AI in this area. This review assesses the role of AI in education, highlighting AI-driven technologies such as automated writing evaluation and natural language processing that provide timely feedback, support personalized learning, and reduce instructor workload. The study employed the PRISMA method to investigate the impact of AI on written English assessment by analyzing empirical studies from Scopus, Web of Science, and Google Scholar. The findings revealed that AI tools significantly enhanced the evaluation of written English by providing reliable assessments, immediate feedback, fostering learner autonomy, and producing assessments consistent with those of human raters, thereby promoting self-regulated writing and increasing access to quality feedback. However, specific concerns about algorithmic bias, feedback clarity, and the diminishing role of human judgment in assessing creativity and discourse were also found. It was also found that educational automation negatively impacts linguistic diversity, critical thinking, data privacy, transparency, and equitable access to AI tools. The review suggests integrating AI capabilities with human experience to balance technological efficiency and pedagogical integrity in the use of AI tools. It recommends developing hybrid assessment frameworks that integrate AI analytics with human evaluation to enhance fairness, accuracy, and comprehension in written English assessment.