<![CDATA[This systematic literature review investigates the integration of personalized Artificial Intelligence (AI) in higher education, specifically focusing on its impact on learning outcomes, knowledge accuracy, conceptual understanding, and associated risks. Following the PRISMA 2020 guidelines, a rigorous selection process from the Scopus database yielded 49 peer-reviewed empirical studies published between 2022 and 2026. The findings reveal that AI-supported personalized learning, ranging from Generative AI and Large Language Models (LLMs) to Intelligent Tutoring Systems (ITS) and advanced Machine Learning architectures, significantly enhances students' academic achievement, engagement, and self-regulated learning skills across various disciplines. However, the synthesis exposes a critical cognitive paradox: improved learning performance and speed do not automatically translate into deep conceptual understanding. Uncritical over-reliance on AI leads to passive cognitive offloading, diminished critical thinking, and vulnerability to AI hallucinations and factual inaccuracies. Furthermore, ethical and institutional challenges, including data privacy concerns, algorithmic bias, and academic integrity risks, present substantial implementation barriers. To maximize pedagogical benefits while mitigating these risks, this study underscores the necessity of a "Human-in-the-Loop" orchestration framework, where AI functions as an educational scaffold while human educators retain primary authority over critical reasoning and ethical technology adoption.]]>
Copyrights © 2025
