Large media corporations. It's a bit frustrating at times. Sometimes when I read the news or try to find something on Google, I can sense an editorial bias or narrative. A lot of times it's subtle, like something being omitted or a tangential issue addressed.
A prime example of this are modern "fact checks". For instance if you wanted to explore if people are arrested for mean tweets in the UK, you can Google "fact check arrested for mean tweets", which brings you to a fact check that is a cherry picked case of an 11 year old getting arrested for "suspicion of violent disorder" and listed as "Misleading". But there are no follow ups, like what is the definition of "violent disorder" or are there any other cases where people were arrested for mean tweets. And since it's deemed as "Misleading", you're left to believe that people are not being arrested for mean tweets, which is just not true.
I feel LLMs allow you to ask follow ups and reason from pure principles much easier than the canned narrative from some giant corporation. Even when you do run up against a narrative, it's pretty easy to poke holes through it with follow up questions.
Of course the reverse is that you can probably find some evidence that feeds into your existing world view. But I think on the whole AI will be a positive impact on culture in this regard