Pedagogical Voice or Algorithmic Authority? A Critical Discourse Analysis of ChatGPT in Language Learning

Abstract

Artificial intelligence-powered tutors such as ChatGPT, Duolingo, and Grammarly are increasingly integrated into language education, reshaping how authority, correctness, and learning are communicated. While these tools promise efficiency and personalized feedback, their discourse carries implicit ideologies that influence learner identity and perceptions of English. This study critically examines how AI tutors construct pedagogical authority and embed language ideologies through their responses to learner queries. Using Fairclough’s three-dimensional Critical Discourse Analysis (CDA), a corpus of 40 AI-generated responses was analyzed across five functional categories. Findings reveal three dominant patterns: (1) pedagogical authority without dialogue, achieved through modals, imperatives, and praise strategies; (2) linguistic homogenization, privileging Standard English and formal norms while ignoring multilingual variation; and (3) neoliberal framing of learning as an efficiency-driven, individualized process. These results highlight the non-neutral nature of AI discourse and its alignment with broader technocratic and neoliberal logics in education. The article concludes with implications for critical digital pedagogy, advocating for reflexive design, multilingual sensitivity, and teacher mediation in AI-supported language learning.

Ähnliche Arbeiten