Exploring Human–AI Collaboration Patterns in Design Problem Solving

• In Event: Human–AI Collaboration for Enhanced Interaction, collaboration and creativity

Abstract

As AI tools become integral to education, understanding how students collaborate with AI on complex, ill structured tasks is essential. We examined 36 postgraduate students’ interactions with a AI chatbot during a design problem solving (DPS) task. Based on the DPS theoretcial framework, dialogue transcripts were coded into six DPS activities and analyzed using Ordered Network Analysis. Three collaboration patterns emerged: Integrated Recursive Strategy, with tight loops among exploration, justification, and framing; Reflective Evaluation Focus, with cycles between exploration and evaluation; and Surface Level Exploration, marked by information gathering with minimal justification or evaluation. These profiles align with existing human–AI collaboration typologies and suggest tailored adaptive scaffolding and process awareness interventions to support diverse learners in AI assisted DPS.

Authors

• Yingxue Liu, The Education University of Hong Kong

• Shurui Bai, The Education University of Hong Kong