A CONTRASTIVE ANALYSIS OF LEXICAL DIVERSITY IN HUMAN-WRITTEN AND AI-GENERATED ENGLISH AND UZBEK ADVERTISING TEXTS

This study investigates lexical diversity in human-written and AI-generated advertising texts in English and Uzbek from a contrastive linguistic perspective. The research focuses on how artificial intelligence influences vocabulary selection, lexical repetition, semantic variation, and persuasive language in advertising discourse. Particular attention is paid to differences between English and Uzbek advertising texts and to the extent to which AI-generated advertisements reproduce or modify the lexical patterns characteristic of human-authored advertising. The study proposes a corpus-based analytical framework involving Type-Token Ratio (TTR), Moving Average Type-Token Ratio (MATTR), and Measure of Textual Lexical Diversity (MTLD), supplemented by qualitative analysis of lexical fields, evaluative vocabulary, promotional collocations, borrowings, cultural markers, and repetitive patterns. Since TTR is sensitive to text length, the incorporation of length-resistant measures such as MATTR and MTLD provides a more reliable basis for comparison. The analysis suggests that human-written advertisements tend to demonstrate greater contextual and culturally motivated lexical variation, whereas AI-generated texts may display more formulaic patterns and recurrent promotional vocabulary. At the same time, AI-generated texts can demonstrate considerable lexical range, particularly when prompted to produce creative or audience-specific advertising. The English and Uzbek datasets also reveal differences related to cultural values, morphological structure, borrowing, and persuasive strategies. The study contributes to contemporary research on AI-mediated discourse, advertising linguistics, and contrastive lexical analysis.