Аннотация
this article considers how generative artificial intelligence may be used to develop the language competence of Grade 10–11 students in Kazakhstan. The discussion focuses on ChatGPT, Gemini, and Grok and takes into account the multilingual character of Kazakhstani schools and the growing use of digital materials in language lessons. A qualitative comparative approach was used. The tools were considered through six typical classroom tasks: writing and revision, reading, vocabulary practice, preparation for speaking, multilingual mediation, and work with visual or multimedia input. The comparison was based on four pedagogical criteria: correspondence with communicative and level-based objectives, usefulness of feedback, suitability for multimedia tasks, and practicality of responsible classroom use. The analysis shows that generative AI may give students additional opportunities to practise, revise, and receive differentiated support. At the same time, its responses may be general, inaccurate, or unsuitable for a particular class. For this reason, the article argues for teacher-guided use and proposes a seven-step sequence that begins with a clear learning objective and independent student work, followed by limited AI support, verification, revision, and reflection. In the Kazakhstani context, this approach also requires attention to unequal access to devices, academic integrity, multilingual classroom needs, and the protection of minors' personal data. The educational value of generative AI therefore depends mainly on the way a task is organised and supervised rather than on the choice of a particular platform.