I think the problem space needs to be divided into two - AI generated articles (which the OP's question is about) and AI generated content/comments. The articles - honestly, normal flagging just works. If the content and the information in it is worthwhile, I don't think people would flag it. Isn't this the whole point of flagging?

However, the second part of the problem - the AI comments, that's really what kills the discussion and eventually the community. If I have trust issues that I'm talking to a bot but not a real user, I am not going to engage further. I might or might not flag the account, but at some point, I'm going to be tired of reporting if everywhere around me it's just bots pretending to be humans. I think this is the more serious problem that needs focus.

In my experience, sloppy AI content almost always, sometime instantly even gets flagged out.