Abstract
<jats:p>Abstract. This paper presents a comprehensive theoretical and analytical investigation into the integration of generative artificial intelligence (GenAI) tools within the higher education system, specifically examining their profound impact on the development of students' critical thinking skills. The central premise of the study substantiates the argument that GenAI can serve as an effective instrument for epistemic development and cognitive enhancement exclusively under the conditions of a purposeful, highly structured didactic design. Such a design must inherently demand in-depth semantic analysis, cross-comparison of diverse academic sources, metacognitive reflection, and rigorous factual verification from the learners. Conversely, the study provides evidence that the uncontrolled, unreflective, and passive application of GenAI technologies inevitably provokes detrimental psychological and educational phenomena. The research meticulously outlines the concepts of «cognitive offloading», «cognitive debt», and the «confidence paradox», illustrating how excessive trust in the linguistic fluency and syntactic perfection of algorithmic systems paradoxically diminishes students' analytical efforts and promotes severe automation bias. To address these complex challenges, the article thoroughly analyzes a conceptual framework illustrating the dual impact of GenAI (DI-GAI-CT), highlighting the dynamic interplay between cognitive stimulators and inhibitors. Furthermore, it details an iterative «human-in-the-loop» interaction algorithm and a multi-component structure of AI literacy tailored for university students. The research proposes and scientifically justifies four specific, actionable didactic scenarios for classroom implementation: utilizing AI as a «devil's advocate» to challenge student hypotheses, conducting blind peer-reviews of intentionally flawed AI-generated texts, performing pre-mortem analyses of project failures, and dissecting the anatomy of algorithmic hallucinations to combat digital misinformation. Finally, the paper identifies the key pedagogical conditions necessary for the productive application of GenAI. These include a critical shift in assessment paradigms — moving the focus from the final submitted product to the transparent, documented process of its creation — and the establishment of clear institutional policies regarding academic integrity. Ultimately, the research affirms that critical thinking is fostered not in opposition to artificial intelligence, but through a methodically guided, highly reflective, and critical interaction with it.</jats:p>