Artificial Intelligence in Medical Writing Education for Non-Native English Speakers: A Systematic Review of Applications, Benefits, and Risks
TL;DR
Across the reviewed literature, AI tools were most often linked to grammar correction, vocabulary choice, readability, writer confidence, and the simplification of medical information for non-specialist readers.
Abstract
This review focuses on how artificial intelligence has been used to support English medical writing among non-native English speakers. The purpose was not to list every available tool, but to identify what the existing literature shows about their uses, benefits, limitations, and risks. Searches were conducted in Google Scholar, Scopus, and PubMed with terms related to artificial intelligence, medical or healthcare writing, and non-native English users. The search first produced 978 records. After duplicate removal, title and abstract screening, and full-text assessment, 24 sources were included in the review. These sources mainly discussed four forms of AI support: grammar and style checking, terminology and lexical assistance, summarisation or simplification, and large language model-based drafting or revision. Across the reviewed literature, AI tools were most often linked to grammar correction, vocabulary choice, readability, writer confidence, and the simplification of medical information for non-specialist readers. These benefits were useful but mostly short-term. Many studies used small samples, brief periods of tool use, or outcome measures that could not show whether learners became better independent writers. It is safer to view AI as a writing support tool rather than a replacement for teachers, supervisors, or writing instruction. Its use in English medical writing education should be guided by clear rules, disclosure of AI assistance, and careful checking factual accuracy in medical text.