Logic unfortunately tends to fail in the real world. Even seemingly concrete ideas like life and death get tricky when dealing with frozen sperm and viruses. Edge cases abound https://en.wikipedia.org/wiki/Chimera_(genetics) People really want stuff like Bayesian statistics to work. But it’s also fails for similar reasons as it depends on accurately updating a model.

The world is filled with models that are simple enough to use and accurate enough to be useful, sadly that doesn’t make them correct.

Censorship seems so simple on the surface, but like everything else the implementation rarely matches the theory. I would argue for disclaimers as a much better option, even if they also fail.