When the Chatbot is Wrong on Purpose: Using GenAI Role-Play to Build Student Engagement and Critical AI Fluency

In this hands-on workshop, participants will explore how customized AI chatbots can be used to increase student engagement while also teaching students to question AI-generated information. The workshop centers on a role-playing game scenario in which students interact with a chatbot non-player character (NPC) that serves as a guide, informant, but one that sometimes provides incomplete, misleading, or invalid information. Rather than treating chatbot errors as failures, the activity uses them as opportunities for students to practice verification, source evaluation, metacognition, and discipline-specific reasoning. Participants will first experience a short version of the role-playing activity as students. They will then examine the underlying instructional design: how to define the chatbot’s role, set boundaries for its responses, embed intentional uncertainty, and build reflection prompts that move students from passive AI use toward active evaluation. The session will include guided discussion on academic integrity, transparency, and ethical use of AI in higher education. By the end of the session, participants will be able to identify at least one course concept that could be adapted into an AI-supported role-play activity, draft a basic chatbot role prompt, design a student-facing verification task, and describe how inaccurate AI output can be used productively to deepen engagement and GenAI fluency.