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

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.