DEVELOPING THE GRAMMATICAL COMPETENCE OF PROSPECTIVE TEACHERS OF UKRAINIAN LANGUAGE AND LITERATURE THROUGH THE USE OF LARGE LANGUAGE MODELS IN THE STUDY OF LINGUISTIC DISCIPLINES

Abstract

The article explores the pedagogical potential of large language models (LLMs) in linguistics-oriented courses aimed at developing the grammatical competence of prospective teachers of Ukrainian language and literature. It is emphasized that the rapid integration of artificial intelligence technologies into educational environments necessitates a rigorous theoretical justification of the principles underlying the formation of grammatical competence through the use of LLMs in the study of linguistic disciplines. The study analyzes contemporary scholarly research on methodologies for developing grammatical competence in future language teachers, as well as on the implementation of digital technologies in education. A typology of instructional tasks employing LLMs is proposed, targeting different components of grammatical competence. These include tasks focused on diagnosing grammatical errors, critically evaluating LLM-generated responses, correcting errors in machine-generated texts, conducting linguistic experiments, and identifying typical syntactic features of artificial intelligence–produced texts. Particular attention is given to research-oriented tasks that involve step-by-step syntactic analysis of texts generated by LLMs in order to identify characteristic structural patterns and compare them with those found in authentic Ukrainian-language texts. The article substantiates the need to foster students’ critical thinking and digital literacy, including their ability to recognize the limitations of artificial intelligence technologies. It is noted that LLMs, whose training data are predominantly English-language, demonstrate lower accuracy in processing Ukrainian, which necessitates mandatory verification of generated content against authoritative sources, such as academic grammars and normative dictionaries. The study concludes that the effectiveness of integrating large language models into linguistics education largely depends on instructors’ ability to methodologically and pedagogically incorporate artificial intelligence technologies into the teaching process, ensuring a balance between innovation and the preservation of the fundamental principles of philological education.