From Dialogue to Deep Thinking: Role-Biased Learning Chatbot Agents to Catalyze Critical Thinking in Collaborative Learning Environments

This study examined the design of a role-biased human–AI learning sequence for management education. Sixty-two undergraduates first-hand engaged individually with chatbots configured to embody distinct corporate roles and decision-making biases. The design then transitioned learners into human–human collaboration, punctuated by structured reflective checkpoints. We traced how insights from initial AI encounters were carried forward, contrasted, and interrogated during group discussion, and how learning was consolidated afterward. The orchestration foregrounded perspective-taking, bias detection, and evidence-based justification through embedded “why” and “how” prompts. The experimental design operationalized structured human–AI engagement as a scaffold for critical, collaborative sense-making in classrooms.