Commercial Artificial Intelligence is booming, leaving U.S. government agencies scrambling to regulate it. The Federal Trade Commission has already found a powerful tool: requiring companies to delete models that use ill-gotten data. Most recently the FTC used model deletion in an enforcement action against Amazon, a move described by the FTC as a warning shot against companies racing to train their models on massive quantities of sensitive data. But can model deletion actually force companies to get their act together… or is it too late to turn back an unstoppable machine?