Though the Turing test – a proper Turing test, I mean, not the pop culture version – traditionally has the human party trying to help the interrogator come to the correct conclusion and the non-human party trying to trick the interrogator into getting it wrong, there’s no rule that says it can’t be the other way ‘round.
Picture a Turing test where the robot is trying to help the interrogator correctly identify the interviewed parties, and the human is trying to trick them into guessing wrongly about which one is the robot.
Imagine how that would play out.