The past five years have seen a rapid rise in the use of artificial intelligence (AI) to support scholarly writing in educational contexts. This scoping review examines literature from 2019–2024 on AI-assisted writing in education. We identify key concepts and synthesize research methods used to study this phenomenon. Prominent concepts include generative AI large-language models (e.g. ChatGPT) and automated writing evaluation (AWE) tools. The review followed PRISMA-ScR guidelines, searching major academic databases (e.g. Scopus, Web of Science) and applying predefined inclusion criteria. A flowchart (Figure 1) summarizes the study selection and categorization process. We summarize representative empirical and review studies in Table 1. Results show that many recent studies focus on LLM-based writing assistants and writing feedback tools. Methods range from qualitative case studies and surveys to quasi-experiments, with a few systematic reviews (e.g. Imran & Almusharraf, 2023) emerging. Common findings include improved writing efficiency and positive student perceptions, alongside concerns about plagiarism and the need for AI literacy. Gaps include a lack of longitudinal or large-scale controlled studies and inconsistent theoretical grounding. We recommend more rigorous designs, reporting of all outcomes (including null effects), and integration of learning theory. This review provides an integrated snapshot of current research on AI-assisted scholarly writing in education and suggests priorities for future work.