The institutional response to AI in education followed a predictable arc: detect, penalize, redesign the prompt. What ended that arc was not a technical failure of detection but a rhetorical one. The detection debate required treating the essay as an object of integrity — something that could be verified as authentically produced. Once it became clear that verification was technically intractable and institutionally expensive, the question shifted from 'how do we protect the essay' to 'what was the essay actually proving.' That is a different and more destabilizing question, and…