Work

Gender Representation in AI-Generated Language: A Conceptual Linguistic Analysis of Bias, Stereotypes, and Representation in Large Language Models

Year: 2026
Type: article
Abstract: Large language models (LLMs) increasingly mediate writing, information retrieval, education, recruitment, and everyday communication. Their linguistic outputs therefore matter not only as technical products but also as forms of social representation. This conceptual paper examines gender representation in AI-generated language by synthesizing evidence from computational studies of bias with insights from critical discourse analysis and systemic functional linguistics. Rather than collecting new model outputs, the paper draws on peer-reviewed research and authoritative reports to identify recurrent patterns in how gender is represented, stereotyped, and evaluated in language-model behavior. The synthesis indicates that gender bias is often realized through occupational association, semantic prosody, emotional attribution, agency, evaluative wording, and recurring assumptions about social roles. The paper proposes a linguistic framework in which representational bias is analyzed across lexical-semantic, grammatical, interpersonal, and discursive levels. It argues that bias evaluation should move beyond counting explicitly stereotypical terms and examine how language choices distribute visibility, agency, authority, affect, and normality across genders. The framework offers a basis for future empirical studies of gendered AI discourse while acknowledging that model behavior remains prompt-, language-, task-, and version-sensitive.
Language: English
Subfield: Gender Studies
Open Access status: diamond
OpenAlex ID: W7220770150
FWCI:
Cites: 9
Cited by:
Related to: 10