Meaningful human oversight: not a rubber-stamp

Human oversight under the AI Act is not about having a person watch a screen and click 'approve' after the system has already made the call. The regulation demands that oversight be genuinely capable of overriding or stopping the AI's output when risks to health, safety, or fundamental rights arise. For example, a medical diagnosis tool that only lets a doctor confirm its recommendation fails the test; the doctor must be able to reject or alter the decision before it takes effect.

To spot exam questions on this, look for wording that suggests oversight is 'automatic' or 'post-hoc'. If a scenario describes a human reviewing a decision after it has been implemented, that is not meaningful oversight. Another trick: remember that the AI Act ties oversight to the specific risk level—high-risk systems require measures like 'stop buttons' or 'human-in-the-loop' designs, not just a log for later inspection. If the answer mentions 'rubber-stamping', it is almost certainly wrong.

A compact way to test yourself: ask whether the human could have changed the outcome before it affected anyone. If the answer is no, the oversight is not meaningful. Keep this contrast in mind—meaningful oversight means the human has real authority to intervene, not just a checkbox to tick.