I don't think people should post the unfiltered output of an LLM as if it has value. If a question in a comment has a single correct answer that is so easily discoverable, I might downvote the comment instead.

I'm not sure making a rule would be helpful though, as I think people would ignore it and just not label the source of their comment. I'd like to be wrong about that.