The fundamental problem of online "discourse" is that critical thinking is applied selectively according to emotional biases, not that participants are intellectually incapable of it. Malicious actors, doubly so. An attempt to control that can only devolve into testing for the "correct bias", and communities with sufficient moderator:post ratio or upvote/downvote type community moderation already perform that sort of self-selection.

As for monetary fees - demographic skew aside, that would significantly reduce the overall user population the more you charge, so a malicious actor would need correspondingly fewer sock-puppets to create the appearance of X% of the population saying Y, limiting its effectiveness as a deterrent. Amplifying minority views by a large factor or seeding new claims/statements doesn't require that many unique accounts in the first place.