Introduction. In the context of the digital transformation of education and the rapid adoption of generative artificial intelligence technologies, students actively use AI tools to complete academic tasks. However, the nature of their interaction with AI remains insufficiently studied. Empirical evidence shows that students’ critical evaluation of AI recommendations varies depending on the activity context and the perceived significance of consequences, indicating that criticality in AI interaction is context-dependent. Purpose. The study aims to substantiate the structure of students’ experience of critical interaction with artificial intelligence and to identify the pedagogical conditions for its development in higher education. Materials and Methods. A comprehensive empirical study conducted at the National Research University “MPEI” included three stages: a survey of academic staff (N = 744), a survey of students (N = 6,328), and an assessment of critical interaction with AI among first-year students prior to the start of their studies (N = 2,898) in situations differing in the perceived significance of consequences. Results and Discussion. A pronounced dependence on context was identified. Critical strategies accounted for 85% of responses in academically significant situations and 44% in everyday situations. Based on theoretical analysis and empirical evidence, a four-component structure of experience was substantiated, comprising cognitive, operational, reflective, and ethical components. Three pedagogical conditions for its development were identified: embedding AI-use situations into learning tasks, organizing reflection on the interaction process, and providing socio-pedagogical support. Conclusion. Developing students’ experience of critical interaction with AI requires its purposeful integration into higher education curricula, which is essential for a successful educational trajectory under digital transformation.