<![CDATA[Artificial intelligence (AI)-based learning tools, including ChatGPT, Grammarly, QuillBot, and Wordtune, have rapidly transformed how students draft, revise, and evaluate academic texts. This article presents a systematic literature review examining the effectiveness of AI-based learning in improving students' academic writing skills. Twenty-five peer-reviewed studies published between 2023 and 2026, drawn from Scopus-, Web of Science-, and DOAJ-indexed journals, were analyzed following the PRISMA protocol. Findings indicate that AI-based learning consistently improves grammatical accuracy, lexical richness, organizational coherence, and revision efficiency across English as a Foreign Language, English as a Second Language, and first-language writing contexts. Motivational and self-regulatory gains were also reported, particularly when AI feedback was combined with instructor mediation. However, effectiveness varied according to students' AI literacy, prompt engineering competence, and feedback type. Overreliance, diminished critical thinking, and academic integrity concerns emerged as recurring moderating factors. This review's novelty lies in synthesizing evidence across heterogeneous AI tool types and educational contexts rather than isolating a single application, thereby mapping both convergent benefits and unresolved tensions. The article concludes that AI-based learning functions as an effective scaffolding mechanism rather than a substitute for writing instruction, and proposes a pedagogical model integrating human mediation to sustain long-term writing competence.]]>
Copyrights © 2026
