Right now I feel like the economics of AI/LLMs are in the spammers favour. You're almost suggesting that AIs be trained to recognise other AIs. Don't misunderstand, I think it would be great if we could opt to have all AI or auto-generated content removed from our internet experience. It's just not really in the interest of the shareholders.

The best thing I've seen in a long time is the "Say potato" tactic against bots, not sure how well it works against LLMs though.