Keyword: «age differentiation»

The article presents the results of a theoretical and experimental study on the effectiveness of using artificial intelligence technologies for correcting lexical collocations among 2nd and 7th grade students. The author, drawing on the classical methodology of E.I. Passov and modern concepts of digital linguodidactics by P.V. Sysoev, substantiates the need for a differentiated approach to the use of AI tools at different stages of education. During a pedagogical experiment (40 students, 6 weeks), the experimental groups used correction of fixed collocations with the help of generative neural networks, while the control groups were taught using traditional methods. The results of the final testing (Post-test) showed that the absolute increase in knowledge in the experimental group of 7th graders was +34% (from 55% to 89%), and in the experimental group of 2nd graders – +18% (from 52% to 70%). The author concludes that AI performs different didactic functions depending on age: a stimulation-motivational function (game form, anxiety reduction) for primary school children and a cognitive-corrective function (collocation analysis, prompt engineering) for adolescents. The article provides practical algorithms for interaction with AI (ChatGPT, DeepSeek, Alice AI) and offers methodological recommendations for integrating intelligent simulators into school practice without the risk of «blind» copying of ready-made answers.