The Paradox of ChatGPT Use in Higher Education: A Systematic Literature Study on Academic Benefits and the Risk of Cognitive Dependency Among Students

Abstract

The rapid adoption of ChatGPT in higher education has produced a paradox between expanding academic benefits and the emergence of cognitive dependence risks among university students. This study aims to identify the dominant forms of academic benefits, map the indicators of cognitive dependence, and analyze how the relationship between these two dimensions has been conceptualized in recent literature. A Systematic Literature Review approach was employed, following the PRISMA 2020 guidelines. The literature search was conducted in the Scopus database for English-language peer-reviewed journal articles published between 2023 and 2025. Methodological quality was appraised using the Mixed Methods Appraisal Tool, and a thematic synthesis was applied to the included studies. The synthesis revealed two opposing clusters of findings. On one hand, ChatGPT was consistently reported to enhance academic performance, learning efficiency, and support for writing tasks and formative feedback. On the other hand, its use without pedagogical scaffolding triggered cognitive offloading, weakened critical thinking, and fostered over-reliance on AI-generated outputs. The relationship between benefits and risks is contextual, contingent upon instructional design and students' AI literacy rather than being intrinsic to the technology itself. These findings underscore the need for institutional policy frameworks and integration of AI literacy into the curriculum to position ChatGPT as a reflective dialogic partner rather than a cognitive substitute.