I think it should matter. When you ask the AI something you are in a frame of mind, you have a specific context, the question also holds value and context that might completely change the parsing of the answer or at least the difficulty of it.

What I'm asking and the response from AI through an intermediary lose some context (the prompt), it's like the telephone game where the data becomes more and more distorted, that's why people don't have an issue with their own AI generated answers.

Another issue is that when I'm talking with someone and parsing through what they've said I'm considering them, as a person, taking all available context (some of this might happen unconsciously).

In any case I don't think there is an easy solution to the problem.