The material difference with the Internet is ROI. If you're going to attack your enemies, the ROI of a troll farm is thousands of times higher than that of a standing army and a conventional military campaign. The ROI of an AI-powered automated troll farm is even higher.

The result is a kind of anti-literacy. Most people can read the words, comparatively few people are media-literate enough to filter truth from lies with any reliability. So the current media landscape is unusually poisonous. It's mostly vested interests lying to you and trying to manipulate you, through ads, troll farms, and mainstream media.

The fix would be AI filtering of content. Right now there's no chance whatsoever of that working accurately, but it's possible in principle to counteract the rise of AI disinformation with AI critiques of it.

Among all of the other revolutionary changes promised by AI, that possibility has flown under the radar. But it would be a political and economic showstopper if implemented, because everyone would suddenly be seeing authoritative, accurate news and analysis - like old-school fairness doctrine journalism, but better because it could presented at a level that matched the reader, while also allowing questions.

Ironically Grok was doing something like this for a while, until it... wasn't any more.