Assessing AI-Supported English Academic Writing in Higher Education: A Narrative Review of Challenges, Emerging Practices, and Prospective Strategies

Abstract

Artificial Intelligence (AI) is reshaping the assessment of English academic writing in higher education, raising questions about validity, fairness, and authorship in AI-mediated learning environments. As students increasingly write with tools such as ChatGPT, Grammarly, and QuillBot, the boundary between human and AI contributions becomes less clear, complicating how assessors interpret performance as evidence of underlying ability. Using a narrative review approach, this article synthesises research on assessing AI-supported English academic writing across three domains: challenges, emerging assessment practices, and prospective strategies. The review shows that ethical and academic integrity concerns, unequal access, algorithmic bias, pedagogical tensions around higher-order thinking, technical limitations in AI-generated text, and institutional constraints on governance and academic staff preparedness are closely interrelated. In response, emerging practices highlight a shift toward process-oriented assessment, including AI-informed feedback systems, adaptive learning tools, and structured AI literacy instruction that makes students’ reasoning and decision-making more visible in assessment. Prospective strategies point toward assessment designs that position AI as scaffolded linguistic support within transparent draft-and-revise cycles, retain human judgement for evaluating complex reasoning and disciplinary communication, and embed clear institutional guidance for ethical and equitable AI use in line with SDG four on quality education and SDG 10 on reduced inequalities. Overall, this review presents an integrated synthesis of challenges, emerging practices, and prospective strategies, showing that credible assessment of AI-supported English academic writing requires process-focused design, clear ethical guidance, and institutional capacity-building across higher education

Ähnliche Arbeiten