I think mirroring free speech standards under the law is probably a good starting point - i.e., censoring is bad by default, but not all speech is protected, an

I think mirroring free speech standards under the law is probably a good starting point - i.e., censoring is bad by default, but not all speech is protected, and calls for violence and insurrection are a relevant and timely example of non-protected speech. I'll reiterate that this should be a starting point, not necessarily a hard and fast rule - there may be more nuanced situations where the standards that are appropriate for the government are inappropriate for social networks. There's also the question of what levels of the stack it's appropriate for any censorship to occur at. Even if censorship of a particular class of speech is appropriate in general, it may not be appropriate for an ISP specifically to conduct that censorship.