Assessing L2 Student Writing in the AI Era: A Systematic Review on Challenges and Best Practices

This paper aims to provide a systematic review of the existing literature on the effects of AI on L2 (ESL/EFL) student writing and assessment. A comprehensive search was conducted through ScienceDirect, ERIC and Taylor & Francis Online, providing qualitative analysis of three major themes: the effects of AI on L2 writing, challenges in identifying AI-generated content, and practices for adapting writing assessment. PRISMA 2020 guidelines and the CASP checklist were used to select appropriate articles and assess validity, relevance and ethical issues. The review included 20 studies published between 2023 and 2025. The results show that AI offers numerous advantages, such as giving individualized feedback and support, improving writing quality and helpfully assisting the grading processes. However, issues related to academic integrity, originality and the difficulties in distinguishing between AI-created writing and student-written work have emerged. The paper proposes best practices for integrating AI into assessment frameworks in which academic integrity is still maintained while AI is utilized to improve learning outcomes.