Experiencing AI Together: A Design-Based Research Approach to Collaborative Prompting in Education

Abstract

Slides of the presentation at ICL2026 (TH Köln, Cologne, Germany). The talk reports on a Design-Based Research study on collaborative prompting in teacher education using the platform collab.prompting.school with in-house LLM hosting. Building on the RETURN Manifesto, it explores how AI use can become observable, discussable and revisable without requiring learners to disclose their full private AI histories. Baseline surveys show that about 70% of participants were reluctant to share their entire prompt history, while they engaged productively with controlled collaborative prompting across four settings. Qualitative feedback shows that prompt revision became central, suggesting that collaborative prompting should make revision pedagogically visible rather than focusing only on final artefacts or best prompts.