And even there, we have made that determination e.g. for speech inciting violence. That's a fairly well-defined case, so I assume that's why we all agreed we're OK with that.
As a thought experiment - or, really, to help me understand - what if we declared "fake news" as "intentionally wrong in substantial facts" and had a very clear definition of substantial, and what counts as intent?
Second thought experiment: What if you could still say whatever you wanted, but "fake news" simply wouldn't be content that's promoted into people's streams? Is that still censorship? After all, you can still say whatever you want - but individuals won't have to hear you when they haven't consented?
(I happen to believe that the core issue isn't the fact that there are fake news, but the fact that the various social media ranking algorithms violate consent by injecting things they deem interesting. I.e. it's not about speech, per se - it's about the fact that the platforms allow third parties to inject themselves into conversations unasked)