Biography
Kaouther Zerakoui is a second year PhD student in English Language and Discourse Studies at Yahia Fares University of Médéa, Algeria, where she also teaches English. Her research interests include critical discourse analysis, systemic functional linguistics, artificial intelligence-mediated communication, media discourse, academic writing, and English language teaching. She has presented at national and international conferences and has conducted research on discourse, ideology, and language in higher education. Her current research explores AI-mediated communication through the lens of Critical Discourse Analysis, examining the relationship between language, power, and meaning in educational and professional settings.
Conference contributions 1
THE ILLUSION OF NEUTRALITY:
A critical discourse analysis of hedging and ideological positioning in ChatGPT
Large Language Models (LLMs) are increasingly promoted as objective and impartial tools that can provide balanced information across a wide range of topics. Despite growing scholarly attention to algorithmic bias in artificial intelligence, comparatively little research has examined the linguistic mechanisms through which AI-generated discourse constructs an appearance of neutrality. This issue is particularly significant in the case of ChatGPT, whose responses are frequently perceived as unbiased despite concerns regarding the ideological assumptions embedded within its training data and output generation processes. Drawing on Norman Fairclough's three-dimensional model of Critical Discourse Analysis (CDA), this paper investigates the role of hedging as a discursive strategy in the construction of neutrality within ChatGPT responses. The study analyzes a corpus of responses generated from prompts addressing selected controversial social issues, with particular attention to linguistic features such as modality, mitigation, evidentiality, attribution, and evaluative language. Through a systematic examination of expressions such as "some argue," "it may be," and "research suggests," the paper explores how ChatGPT negotiates competing viewpoints while simultaneously positioning certain perspectives as more legitimate, credible, or authoritative than others. The study argues that neutrality in AI-generated discourse should not be understood as the absence of ideology, but rather as a discursively constructed effect achieved through specific linguistic choices. By foregrounding hedging as a central mechanism in the production of apparent objectivity, this research contributes to emerging discussions on language, ideology, and algorithmic meaning-making. More broadly, it demonstrates the relevance of Critical Discourse Analysis for understanding the social and epistemological implications of generative artificial intelligence in contemporary digital communication.