Practising and problematising generative AI: Higher education students’ discourses, ambivalences, and the challenge of thinking (with) a cultural technology

This study examines how university students engage with generative artificial intelligence as a cultural technology within higher education. Drawing on 40 reflexive group diaries produced during a university course and analysed through thematic analysis, the research explores the discursive figures through which students negotiate authorship under algorithmic conditions and engage in processes of situated subjectivation. We identify five symbolic figures (Ally, Obstacle, Interlocutor, Routine Presence, and Threat) reflecting distinct modes of relating to artificial intelligence. Analysis of prompt design, editorial adaptation, information retrieval, and genre simulation reveals how students perform curatorial labour and exercise agency within algorithmic frameworks. We argue that artificial intelligence operates as a dispositif of soft algorithmic discipline, normalising particular stylistic patterns and eliciting reflexive self-regulation. Students deploy moral distinctions, disclaimers, and emotional framings to legitimise their artificial intelligence use, suggesting that ethical engagement is integral to technological adoption. We propose two conceptual constructs: editoriality under algorithmic conditions , which captures the curatorial and hybrid mode of artificial intelligence–assisted composition; and performative para-intelligence , which describes the pressure to conform to algorithmically mediated standards of academic expression. These findings advance a cultural sociology of generative artificial intelligence attuned to the ambivalent, reflexive, and situated ways in which users negotiate authorship, legitimacy, and identity.