Prompting as Discursive Practice: A Critical Discourse Analysis of Hypothetical Prompts in the Context of Nigerian University Students' Use of Generative AI

This study examines prompt construction as an emergent digital genre in the academic use of Generative Artificial Intelligence (Gen AI) among Nigerian university students. Drawing on Teun van Dijk’s (1993) socio-cognitive model of critical discourse analysis (CDA), it argues that prompting should be understood not as a technical means of eliciting information from an AI system, but as a discursive practice through which assumptions about knowledge, authority, language, and cultural context become consequential. The paper is conceptual, not empirical, and does not analyse a collected corpus of student prompts. It develops a framework from van Dijk’s account of discourse, cognition, and society and illustrates its potential through hypothetical prompts modelled on patterns reported in scholarship on Nigerian students’ use of generative AI. Six constructed prompts serve as the hypothetical data. Proposed dimensions include: epistemological framing, linguistic and contextual absence, and mood, modality, and authority positioning. It argues that prompt construction can operate as a form of pedagogical power, as routine prompt conventions may reproduce epistemic hierarchies without deliberate intent. It concludes by advocating AI literacy that extends beyond technical competence to critical interrogation of AI systems’ epistemic assumptions and proposes a corpus-based research agenda for testing the framework.