There should be multiple independent steps of verifying if someone should get banned and in what way. And probably integrate a good test for joining the community so that it’s more likely for people to be rational from the start (that way you don’t even have to look at so many potential flags).

How much would you pay to join a community with that level of protection for user rights? Like the old subscription based forums, some of which are still floating around the internet?

Because “multiple independent steps of verifying” is, frankly, going to be a lot of frustrating, thankless, and redundant work for moderators. I mean, we know how to safeguard people’s rights through legalistic processes. Courts do it all the time. It’s called due process. And due process is frequently a slow, complicated, and expensive pain in the ass for everyone involved. And I think very few people would want to do that work for free.

(Conveniently, this would also serve as a good test for joining such a community - people are more likely to follow the rules and act like decent human beings if a subscription they paid for is riding on it, and it would price out AI and spambots in the process.)