The reasonable doubt standard does require the benefit of doubt go towards innocence. Presumably the goal is for more guilty to go free than innocent being punished.

Which is different standard. And it applies only in actual trial, which is something quite rare and super expensive. Most of cases are decided outside of trial, where going to trial raises stakes a lot.

It isn't the goal as much as an unavoidable effect. I however don't feel that comparing video removal to incarceration is relevant. Compare it to traffic rules instead (or anything else you can think of). You have to stop unnecessarily at times, and can't move faster than the speed limit, etc, just because we have deemed the fail state worse than the needless limiting of good drivers' freedoms.

I'm not saying which is the case here, but certainly that the argument could be made.

> It isn't the goal as much as an unavoidable effect.

I don't see the value of this distinction. Creating a system where it is more likely that guilty go free is the practical equivalent of creating a system where more guilty go free.

I wasn't commenting on how well the analogy fits censorship. Though I would note that the criminal justice system doesn't set all people free, just that it adopts a more slanted standard of proof than the laxer standard used in some other judicial situations like civil and immigration courts. The discussion over allowable levels of false positives fits better than a dichotomy comparing a world with traffic laws to anarchy or something.