STYLISTIC FEATURES OF ENGLISH TEXTS GENERATED BY ARTIFICIAL INTELLIGENCE

Abstract

The article examines stylistic features of English texts generated by artificial intelligence (AI) and provides a comparative analysis with texts authored by humans. The main focus is on identifying characteristic markers that allow differentiating machine-generated texts from human-written ones, as well as determining the specificity of formal neutrality and academic standardization in AI texts. The study revealed that AI texts are characterized by a high level of formality and stylistic neutrality, manifested in the use of academic vocabulary, repetitive clichés, passive constructions, and complex subordinate syntactic structures. At the same time, these texts demonstrate low variability of syntactic and lexical means, lack an authorial voice, and have minimal emotional coloring, making them universal, grammatically correct, and logically structured, yet less informative and stylistically limited compared to human texts. At the lexical level, the analyzed texts show a dominance of abstract and generalized academic vocabulary, with limited use of emotionally colored words, metaphors, and idioms. Syntactically, AI texts typically feature long complex sentences, parallel constructions, and frequent use of passive voice, while inversions, parenthetical remarks, and expressive elements typical of human writing are absent. Discursive analysis indicates that AI employs standardized transitional phrases and repetitive syntactic patterns while avoiding controversial or subjective statements. Comparison with human-authored texts shows that humans use more diverse lexical and syntactic means, authorial assessments, emotional intonations, and stylistic variations, providing individuality and expressiveness to their writing. The results indicate that stylistic markers, levels of neutrality, formality, and standardization of syntactic and lexical structures can serve as reliable criteria for identifying AI-generated texts, which is practically significant for academic, educational, and professional contexts. The study opens prospects for further development of methods for evaluating the quality of machine-generated texts, improving authorship recognition systems, and studying the impact of AI on contemporary English communication.