There’s no reason such a contextualization engine would need a singular source of authoritative information, rather I think it would be useful as a tool to be c

There’s no reason such a contextualization engine would need a singular source of authoritative information, rather I think it would be useful as a tool to be configurable against the sources that the reader trusts or a set of sources backed by an organization the reader trusts. One key thing would be to adjust the the authoritative-ness of different sources based on the local context of individual parts of text. For instance a paragraph of an article might have references to psychology and to history, I would generally trust very different people to give authoritative information about those topics. This gets trickier when dealing with topics that blur the boundaries or combine different fields. Obviously you could just show more and more information but at some point that overwhelms the reader. In any case this would not solve misinformation, but a key thing it would help with is dumb misinformation. Little fibs that could be easily disproven if checked but the reader does not have any indication to do that checking. For instance someone claims to be a professor at Harvard, let’s look that up. Ideally the frequency of the dumb misinformation can clue the reader to the fact there might be some smart disinformation there. But sometimes people are careful about the simple stuff in disinformation, and sometimes people are writing honestly but are just wrong.