Designing GenAI as Dialogic Partners: From Pedagogical Principles to Educational Chatbots

In this seminar, I present a theoretical and pedagogical approach to GenAI as dialogic partners in learning. Rather than treating GenAI primarily as a source of information or feedback, the approach focuses on how it can be designed to participate in forms of interaction that stimulate reasoning, questioning, and the development of ideas. Drawing on an ongoing research programme, I show how dialogic principles can be translated into the design of no-code customised educational chatbots through three cases focused on argumentation, multimodal reasoning, and translanguaging. I also illustrate how student–AI interactions can be analysed to examine whether and how these intended pedagogical principles are enacted in practice. The seminar uses these cases to consider broader questions about pedagogical design with GenAI: How can educational purposes be translated into chatbot design? What forms of dialogue are productive for learning and assessment in different contexts? And what are the possibilities and limits of positioning GenAI as a dialogic partner?