I reckon AI has it's uses - things like generating images to illustrate human written posts provided they are credited in the normal way, perhaps for translating posts from one language to another in a way that handles slang and phrasing well.
It's a shame that AI-generated text doesn't self-identify. I wouldn't have a problem with a post which clearly identified as written by AI, and stated the seeds used. That way, I could decide for myself whether it was worth an upvote or not (I can see some situations where it would be).
It's when AI-generated text is posted as if written by a human that I have a problem, then it most definitely is fraud. Luckily, I'm starting to see an "uncanny valley" kind of effect, where there is something subtly different about AI text that I can't put my finger on, but I can sense it is just wrong.
I have no problem with such use cases. The problem is with users putting out ai generated content and pretending it is theirs.