Whether the meme is meant to be shared in some other context or not, I think, is the decision that should be based on the sum of copyright liberation and how generalistic contents are. Today, I can't bear a thought of reposting some stranger's niche meme on social media without attaching a source/creator, because I'm most likely making one more point where engagement with the same meme ends - and reposting full works doesn't qualify as commentary/criticism/research, so it's not a fair use, to begin with.

That's why we are correct of assuming the worst from the bots: programming any fair use considerations is left to gather dust, as it's ultimately left for human to decide such things.