Currently people can create such information, of course. But liars routinely deny well-documented things in their past, and those denials are effective to their intended audience. And currently someone can debunk a lie, but there's no systematic way to find everyone who testifies that it's untrue.
And while finding all of a claim's supporters and debunkers would be a great first step, a truly good (i.e. skeptical) reader would want to be able to find their arguments, too.
I'm skeptical that current AI could automate the job of indexing that information -- understanding controversial arguments seems likely to require true intelligence. I happen to have made a system[1] in which indexing any kind of data or metadata is extremely easy, requiring little more than knowing some natural language. (Deciding what kind of information you want to encode is hard, but once you've decided that, encoding it is easy.)