The article is devoted to analyzing the psychological impacts of artificial intelligence (AI) tools on school students in the context of the global spread of AI and its active integration into the educational environment. The aim of the study is to systematize the positive and negative psychological effects of AI use, identify factors contributing to the development of dependency and cognitive overload, and determine prospective research directions for minimizing risks and optimizing the application of AI in education. Methodologically, the work is based on a review of contemporary literature and a quasi-experiment conducted in real classroom settings.

The results indicate that AI use with pedagogical support is associated with several positive effects: increased learning motivation, personalized learning trajectories, timely feedback, and the development of self-regulation and metacognitive strategies. At the same time, potential risks are also observed: cognitive overload under high intensity and multitasking, attention fragmentation, reduced independent effort in problem-solving, development of behavioral dependence on «prompting» assistance, increased anxiety/uncertainty, as well as ethical and privacy threats. The emergence of these risks is influenced by students’ age and digital literacy skills, the frequency and duration of AI use, the type of learning tasks, interface patterns (notifications, auto-completion), and the lack of transparency and pedagogical regulations.

Practical implications include recommendations for implementing «distraction-free» modes, structured pedagogical mediation («teacher-in-the-loop»), explicit labeling of AI assistance, time and task difficulty norms, and AI literacy programs for teachers and students. Prospective research directions include longitudinal assessment of «dose-dependent» effects, experimental testing of interface «frictions», development of valid metrics for cognitive load, and protocols for ethical review in school practice.