In what sense will it kill social media?

Unless social media websites begin to require ID verification, they'll be flooded with LLM-powered bots. Humans will tire of this.

We can't be far off the state where any captcha-style humanity test that can be implemented is immediately vulnerable to bots, too.

Not captchas, OP probably means actual IDs. And to be honest, with some cryptography surely we can find a way to make it work and be anonymous

I simply cannot imagine being willing to provide identity documents to any of the social media/advertising companies. They would have to somehow demonstrate they're trustworthy first.

The bit of social media people use to talk to their IRL friends and family will be fine.

But the bit of social media where people are listening to celebrities, getting their news, sharing memes, discussing common interests? It might get flooded with bots so convincing as to be indistinguishable from normal users.

It's as soon as it learns that X celebrity is dead generates clicks.

Then folks will learn not to trust it just because it is a computer once everyone is announced dead, and everyone will have to make claims that they aren't.

So Facebook is the one most likely to survive. Reddit, Twitter, Hacker News, ... etc. soon to be flooded with LLM generated spam.

Last I checked, non AI AdBlock already works pretty well, why would you need AI?

Why do you believe that users will start using this "AI AdBlock"? Regular ad blockers are enough to cover current ads. Google also controls the main browser in the biggest mobile OS (Chrome on Android), and it doesn't support ad blockers. There are options (Firefox, Brave, Kiwi, etc.), but most users don't bother.