The rapid diffusion of generative artificial intelligence (GenAI) tools such as ChatGPT, Gemini, and Copilot into higher education has raised urgent questions about their influence on university students' critical thinking. This paper reviews and synthesizes recent empirical and theoretical literature to examine how GenAI shapes analytical reasoning, evaluation, and reflective judgment among students. Drawing on Bloom's revised taxonomy and the cognitive-offloading framework, the paper argues that GenAI's impact on critical thinking is neither uniformly positive nor uniformly negative but is contingent on how the technology is used, the learner's AI literacy and self-efficacy, and the pedagogical structures surrounding its use. Evidence indicates that structured, reflective engagement with GenAI can strengthen argumentation, evaluative reasoning, and higher-order cognitive skills, whereas passive, uncritical reliance is associated with metacognitive laziness, cognitive inertia, and diminished independent reasoning. The paper further discusses academic integrity concerns and offers pedagogical recommendations, including AI-literacy instruction, process-oriented assessment, and scaffolded prompt-based tasks, for cultivating critical thinking in an AI-saturated learning environment. The paper concludes that deliberate instructional design, rather than the technology itself, determines whether generative AI becomes a catalyst for or an obstacle to the development of critical thinking in higher education.