All of those subjects are just topics: a simple Bayesian classifier can identify those kinds of stories, arbitrating "the truth" (or lack-thereof) doesn't come into this. If empowered with this kind of filter I can assure you that I'd block-out stories about whoever this year's disgraced musician is, or whatever Prince Andrew's getting up to now, regardless of whether they're true or false.
It's not for society, it's for the author's specific needs. On Twitter, you can filter certain keywords you don't want to see. Why not build a tool to do that sort of thing for the entire web?
Yes, let a big button to decide what are facts, what are opinions and what are conspiracy theories for us. Exactly what this society needs.