Keyword: «chain-of-thought»

The article examines the process of transforming skills in interacting with generative neural networks into an independent professional competence within the field of linguodidactics. The study is motivated by the strategic objectives of the national "Data Economy" program, which envisages large-scale integration of artificial intelligence technologies into pedagogical practice. The author traces the methodological development of query construction approaches – from basic instructions to complex reasoning architectures – and systematizes the structure of prompt engineering skills, highlighting linguistic, cognitive, technological, and strategic aspects. The results of an empirical experiment involving linguistics students (a sample of 48 participants), covering diagnostic and formative training stages, are presented. It is confirmed that targeted training significantly improves the quality of educational materials generated by algorithms, demonstrating a 63% efficiency increase based on final assessment results. The feasibility of institutionalizing prompt engineering as a separate academic discipline in the foreign language education training system is substantiated.