The existing literature on the role of generative artificial intelligence (GenAI) in education has mostly focused on accuracy, ethics, academic integrity, and the role these technologies might play in teaching and learning. However, crucial to these discussions is the need to explore their linguistic “choices,” an area still largely unexamined in connection with education. Drawing on a program of research on AI-generated feedback and AI translation set within higher education, I argue that linguistic analysis is essential for understanding how GenAI constructs agency (who is represented as acting or responsible), evaluation (how attitudes or judgments are encoded), stance (the positioning or perspective taken toward information), and coherence (the linguistic mechanisms that create logical flow). These discursive aspects carry significant pedagogical implications, as they can influence how learners interpret AI output, how feedback supports or undermines self-efficacy, and how disciplinary norms are represented. In this article, I posit that linguists can play a major role in dissecting AI discourse, informing calibration, and designing AI literacy pedagogies that can help students engage critically and productively with GenAI.
Gabriela C. Zapata (Tue,) studied this question.