Towards a Theory of Generative AI in Problem-based Learning

This article presents a design experiment in which generative artificial intelligence (GAI) was topically integrated into a Problem-based Learning module in a pedagogical study programme with the intention to generate insights for both, future GAI-in-PBL practice and theory. Based on various data (student reports, notes, a focus group interview, transcripts of teachers/researchers’ discussions), three design elements were assessed regarding their potential to help students to develop their own learning practices with GAI, and regarding how the emergent practices and dynamics enrich existing PBL theory. The analysis revealed students’ weariness, shame and fear for/of using GAI, but also how the PBL process enabled them to develop their own reflective and nuanced GAI practices built on their own learning, integrating community, communication and trust. The experiment also revealed co-knowledge construction amongst students, while stakeholders and teachers were seen as figures of authority on how to approach GAI.