I disagree with the rule banning AI content. Especially the dumb original imgflip AI can just generate super funny dumb captions to classical meme templates, and there is no harm in doing this (you cannot even decide with 100% certainty whether a human or an AI has generated it).
I also used to generate dumb nonsense jokes with the early open source LLMs (long before ChatGPT got released).
I think saying AI content is exploitative in nature is a bit short-sighted and just generalizing about all the different types of AI. If I, e.g. program my own AI to e.g. play a game, and make a meme about that, that's not exploitative. Imitating people's art without their consent is something different though.
I think it would be good to revisit this rule and maybe specify what exact content are not allowed or think about this, since generally there can be good AI meme content (also on a meta-layer, e.g. about people who take obvious fake AI pictures as something real) and just outright banning all of them is a bad idea in my opinion.