Build, Play, Publish: Collaborative AI Game Design for Music Education

Generative AI tools are both extractive and productive (Crawford 2023; Johnson & Salter 2025). Their use in universities is contentious, with some students boycotting AI-related courses (such as those at Berklee College of Music) or joining a renewed Luddite movement. AI enthusiasts, on the other hand, risk offloading not only their intellectual labor but the benefits of that labor to a chatbot. This presentation finds a middle ground by exploring the playful capacities of both proprietary AI platforms (here, Claude) as well as those of publicly available models. We consider collaborative game design as a form of “critical making” (Ratto 2011; Johnson & Salter 2025) that can engage students in the study of music while also developing critical AI literacy.

With this conceptual grounding, our presentation demonstrates how to collaboratively develop games. We show interactive browser-based games that were built with Claude for music courses, and as a group we will build and publish a game together live. We will also show how to use specialized models to generate music and text-to-speech audio that can be used for gaming, research, and teaching purposes. Overall, our discussion will incorporate humanistic and computer science perspectives on the technical development of such models and the power dynamics that govern how they are built and used. Participants leave with a shared published game and reusable workflow.