‘Whack the prompt in there and see what it says’: making sense of Communicative AI in education

Using a communication lens, this article shows how interactions with a persuasive, but not entirely trustworthy, artificial partner (ChatGPT) are explained and ‘made sense of’ in an educational setting. The argument draws on a year-long qualitative study in a language school for international students. The findings draw attention to communicative negotiations about contextual criteria of authority, pedagogical meaning, and trust. In particular, the fieldwork highlighted three distinct dynamics: (a) retrospective rationalisations of AI use and misuse shaping teachers’ sensemaking; (b) simple forms of tactical prompting to accomplish basic tasks among students and (c) a moment-to-moment interplay of communicative roles, accompanied by situational variations in trust attribution. These findings suggest that a focus on pragmatic communicative interactions is a productive avenue of empirical research in education technology research. In the conclusion, the article reflects on the possibility that the analysis of projections, expectations, and desires flowing from human agents to communicative AI systems – and vice versa – could encourage the creation of educational practices of human distinctiveness.