Dampak AI Generatif (LLMS) terhadap Keterampilan Menulis dan Integritas Akademik: Tinjauan Literatur Sistematis

Abstract

This study aims to examine the impact of Generative AI (LLMs), such as ChatGPT, on students' writing skills and academic integrity through a Systematic Literature Review (SLR) of 15 relevant articles. The synthesis results reveal a significant duality effect, creating a dilemma between efficiency and quality. On one hand, LLMs are proven to be effective assistive co-pilots, which clearly enhance the technical efficiency of writing, accelerate the research workflow, and improve text cohesion and precision. However, this convenience triggers a substance quality crisis because students experience cognitive over-reliance, leading to Academic Deskilling the loss of the ability to independently practice critical reasoning and idea synthesis. This dependency is exacerbated by blind reliance on AI output and vulnerability to hallucinations that threaten the originality of the work. The consequence is a shift in the form of misconduct to the sophisticated practice of prompt engineering. To address these challenges, the literature emphasizes the necessity of pedagogical redesign and assessment redesign, which must focus on establishing ethics that ensure full transparency and accountability from students, and designing tasks that demand personal synthesis and reflective thinking.