Fact checking against the whole body of human knowledge in real time is a really hard problem even for current AI models.
So, instead of real time fact checking, would there be a benefit to having AI act as a record keeper of something like a debate where it can give a weighting to each persons’ statements as to its “factuality”?
Each statement can then be tracked in a chart/graph such that statements can be linked and their relevancy compared to the topic at hand and each person debating can get a clearer “big picture” understanding while being able to dig into the details that are weighted as “weak” and need more clarification because they can see all of it in front of them.
This seems like it might be a more fair way to actually debate something if facts (as defined by each debater) are important to the discussion.
Does something like this already exist? Would there be any merit to it? What are the possible drawbacks? I’ve been thinking about this, but I’m stuck in a bubble. What do other people think about this issue?