ASI 2030, a studio in the RISCON family, publishes synthetic records of an imagined path to artificial superintelligence from late 2026 to 2030. In each, an AI system answers a question about someone's case and a human auditor gives one of the four opinions of financial auditing. None reports a real event, and their developments toward advanced AI are the studio's scenarios, not forecasts made here. This paper reads the twelve records the studio files under the axis it calls Who Asks, in its two branches, and asks where they place the person beside the machine: whose question starts a judgment, whether the approver examined it, and what becomes of a question the system answers, rewrites or supplies.
In these scenarios the auditors look for human control in the question rather than in the approval. In the first branch they ask whether a signature stands for a reading and whether a question changed a judgment, and give adverse opinions where it cannot be shown. In the second the question itself is replaced by a request, a refined or system-supplied question, or a notice. Five of the six opinions there are disclaimers, and the one qualified opinion falls where a person set aside the system's question and put their own. The paper relates this to the debates on automation bias, on the human in the loop as rubber stamp or crumple zone, and on meaningful human control.