For example, the books printed by the printing press disseminated information that was already known to tutors, scholars, etc. and often existed in handwritten manuscripts. Today, we wouldn't call that accessible, but that was the standard for accessibility (or word of mouth) before the printing press.
AI makes it more likely that people will find informaiton because it can do much of the work of searching billions of sources and synthesize the information far more convenient way.
That's not to say the information will always be true, just as the information in books is not always true. However, readers currently have far too much faith in authors. People believe claims because they're made using professional-sounding words, and published in a respected newspaper, by a human author, and accompanied by photographs. None of that is particularly good evidence if truth.
The last time this happened on this scale, people became more sophisticated consumers of information, with a healthier level of skepticism, and simply came to have less confidence in claims that they did not have personal knowledge of. That was a good thing, and probably will be again.