A Study on the Influence of AI-Assisted Writing on the Style of English Academic Discourse

Abstract

The impact of AI-assisted writing on the style of English academic discourse has become a focus of academic attention. Through multi-layered linguistic analysis, this study systematically explores the stylistic representations of AI-generated texts in terms of lexis-grammar, rhetorical structure, and human-computer interaction. The study finds that, at the lexical level, the texts exhibit a concentration of high-frequency academic vocabulary and a reliance on nominalization structures. At the syntactic level, they demonstrate a tendency towards formulization and a preference for the passive voice. In terms of discourse cohesion, the texts rely on explicit logical markers, yet exhibit limitations in long-range coherence. Regarding rhetorical structure, AI can achieve standard move progression; however, its identification of research gaps remains superficial. In human-computer interaction, stylistic negotiation and disciplinary conventions shape the heterogeneous nature of the blended texts. This study provides a theoretical framework for understanding the discourse style of AI-assisted writing and offers implications for the teaching of academic writing and technological optimization.