Looks like we are pretty close for someone being able to flood internet discussions with billions of alt-right/far-left/whatnot fake bot accounts. Bad times for moderators and anyone trying to engage in decent discussion.
Less doom and gloom view: Many of the posts here were about humour. That is a terrific use case for AI language models. You can have something like a Medium blog (for monetisation) where the author tries to induce funny responses from the AI langauge model. Best responses become the day's blog post. I would definitely read it. The one posted here about Lord Byron rhyming about SpongeBob SquarePants was hysterical to me. Or what if you asked the AI language model to generate a fake interview between Lord Byron (or Mark Twain) and Lex Fridman. Probably the result would be hysterical, and Lex would love it!