You could make the case that it could be the opposite: people could use AI to confirm what they already believe. Already I think the problem is what people want to consume, for instance my understanding is that the most popular part of the New York Times is the opinion page which is cheap to produce and mostly the likes of David Brooks and Charlie Blow blowing it out their ass. The real news is expensive to produce because it involves sending young reporters out in the the field and do research, sometimes going into harm's way, applying elbow grease and data analysis and in the end it is about some people you don't know in some town you never heard of and doesn't come to a clear moral conclusion.

There is the criticism that the media is harmful and 'biased' because of the interests of the people who own it (conservative) or who work in it (liberal) but a more fundamental criticism that says it gets things wrong by its very nature which includes the works of Neil Postman, Marshall McLuhan, Norman Mailer, Joan Didion, Ben Bagdikian [1], Johnathan Schell, Hunter S. Thompson, etc.

I've lived long enough that I've had the experience that an "event" [2] takes place -- I see it live on TV, then the recap on the evening news, then in the daily paper the next day, then in a weekly paper like The Economist, I read more articles about it, and I think I know what happened.

25 years later a book comes out that is very well research and they combed over the primary sources and interviewed lots of people who were involved and it seems to me that what happened is very different from what I remembered, some of that is my memory being fuzzy [3] but some of it is that journalists write the first draft of history and don't get it quite right?

So I think that's what you're up against.

We're also up against something new in the social media age in that you're aware of the editorial function of the TV channel or newspaper to some extent (but nowhere near as much as you should -- because 10,000x as many things happened today and did not get in the news as did get in the news) but when you see something in social media you think other people care about it as opposed to the editors. Ezra Klein was speculating in his last podcast, for instance, that if Elon Musk was at war with Donald Trump he might amplify Xes about the Trump-Epstein connection which could be devastating because it creates the appearance (and reality) of a movement.

[1] who pointed out the conflict between the owner and journalist mentioned above

[2] l’événement in French theory. Since everything else is under-translated in French theory I think this word, used technically by various authors, ought to be dragged into English

[3] If I was an LLM would we call this "hallucination?"