This paper explores how Donald Schön's concept of generative metaphor can serve as a conceptual frame for problematising AI in education. Our focus is on the metaphor of AI as autotune for knowledge, as originally proposed by educator Dave Cormier. Through a close reading and exploration of the metaphor, we draw on Schön's conceptual materials to surface underlying assumptions and reframe key issues at play in Cormier's story. We show that the autotune metaphor, while intuitively appealing, oversimplifies the analogy between autotune as a tool for professional music-making and the relationship between expertise and use of AI tools. Drawing on a more nuanced view of autotune in music production, we elicit several points of conflict with Cormier's original metaphor to generate a richer problematisation of issues posed by AI in education.