What about "check if the user is right"? For thinking or agentic modes this might work.

For example, when someone here inevitably tells me this isn't feasible, I'm going to investigate if they are right before responding ;)