#166: Biased Chatbot Interaction as Preparation for Collaborative Learning

Twenty-five teachers engaged in individual preparation with chatbots deliberately designed to exhibit biased stances on the learning topic of AI adoption in higher education before engaging in collaboration with their human peers. Through intermediate reflections, we probed into how everyone consolidated their learning from the initial human-AI collaboration, whether and how this affected reasoning during groupwork, and how everyone consolidated their learning post human-human collaboration. Results showed that while teachers recognized and critiqued biases of the chatbots, their group discussions deepened understanding of ethical concerns and the complexity of biased views, leading to a more nuanced and criteria-driven endorsement of AI for education. Overall, our design fostered reflection and perspective-taking, although some teachers remained resistant to changing their views based on feasibility concerns.