<![CDATA[Generative artificial intelligence (GenAI) has rapidly altered academic writing in higher education by expanding access to feedback, idea generation, paraphrasing, translation, synthesis, and text revision, while also destabilising inherited assumptions about authorship, assessment, source use, and academic integrity. This scoping review maps recent journal literature from 2021 onward, anchored in a Scopus export searched on 22 February 2026. The review indicates that, under structured pedagogical conditions, GenAI may improve the surface features of student texts, support formative feedback, and reduce some barriers for multilingual writers, yet these benefits are contingent on instructional design, disciplinary context, task type, students’ language proficiency, and the presence of human verification. Some quantitative studies report substantial gains from structured interventions, including large effects in AI-enhanced goal-setting and improvements in writing quality in anti-plagiarism literacy programs, whereas qualitative studies show more ambivalent changes in writer agency, authorial voice, and metacognitive engagement. Across the corpus, academic writing is being reframed from a solely individual textual product to a distributed, processual, and ethically governed practice involving students, teachers, peers, writing centres, institutions, and algorithmic systems. The review argues that the central challenge for higher education is not whether GenAI should be permitted, but how writing pedagogy, assessment, and policy can preserve intellectual labour while teaching students to use AI critically, transparently, and responsibly within transparent and process-visible writing practices.]]>
