LARGE LANGUAGE MODELS AS A TOOL FOR MEDIA COMMUNICATIONS: A PHILOLOGICAL ANALYSIS OF THE SYNTACTIC AND STYLISTIC FEATURES OF GENERATED TEXTS

Vasilev Andrei Mikhailovich

The article examines the syntactic and stylistic features of texts generated by large language models (LLMs) in the context of their application in modern media communications. Based on the analysis of news, journalistic, and advertising texts generated by GPT-3.5 and other LLMs, the article identifies the characteristic features of automated content, such as the standardization of syntactic structures, limited vocabulary variation, and a neutral-formal style. The study demonstrates that, despite their high literacy and coherence, generated texts retain a template-like nature, which affects their perception in the professional media environment. The results of the work can be used to develop criteria for assessing the quality of automated content and its integration into media communications.

Keywords:large language models, automated text creation, media communications, philological analysis, syntax, stylistics, GPT-3.5, digital linguistics.