<![CDATA[The rapid advancement of Artificial Intelligence (AI) has transformed scientific writing practices in higher education by providing innovative support for literature exploration, content generation, paraphrasing, grammar correction, and manuscript refinement. Despite the increasing adoption of AI-powered writing platforms, comprehensive comparative studies examining their strengths, limitations, and educational implications remain limited. Therefore, this study aimed to investigate the roles, capabilities, and challenges of six widely used AI tools Perplexity.ai, Braintext, QuillBot, Claude, Gemini AI, and Grammarly in supporting scientific writing. This study employed a Systematic Literature Review (SLR) approach following the PRISMA 2020 framework. Relevant studies published between 2019 and 2025 were retrieved from Google Scholar, Scopus, ScienceDirect, SpringerLink, and Taylor & Francis Online. The initial search identified 126 records. After removing duplicate records, 108 studies were screened based on titles and abstracts. Subsequently, 78 full-text articles were assessed for eligibility, and 35 peer-reviewed studies met all inclusion criteria and were included in the final synthesis. The findings revealed four major themes: (1) efficiency and productivity enhancement, (2) improvement of writing quality and linguistic accuracy, (3) ethical and academic integrity concerns, and (4) pedagogical implications for higher education and English education. The review demonstrated that each AI platform offers distinct advantages in supporting scientific writing; however, no single tool comprehensively addresses all writing requirements. Furthermore, challenges related to hallucination, factual accuracy, transparency, and overreliance remain significant concerns. The study concludes that AI should function as a cognitive support tool rather than a substitute for human intellectual engagement.]]>
