From what I’ve seen in your replies, you seem to agree:
Bad actors can easily ruin a community
It’s very easy for bad actors to game popularity-based systems like downvoting posts to remove them or upvoting posts to protect them
Bad actors can brigade communities to make it seem like active members support values different than what the majority actually held before the brigade
You’re dancing around the solution but refuse to admit it: you need a group of trusted users who have a longitudinal relationship with the community. This group of users can follow the community’s leanings over a long period of time, keep the discussion true to the community’s original vision, and easily identify bad actors. You need moderators.
It seems you’d be in favor of more laissez-faire moderation, but there’s still no better solution than moderation. Even if AI got good enough to do the job as well as a human, you’d still need a leader (the community creator or mods) to program the parameters of that AI. The truth is that your anarchist belief system simply doesn’t work as well in practice as it does in theory, and the only viable solution involves having someone in charge.