I think this is the central problem. Companies are overcautious in their assessments because of potential litigation and they also tend to use algorithms for them to save costs, which almost always leads to algorithmic discrimination.
In contrast to this, judges are generally able and trained to make such assessments by evaluating the circumstances and the intentions behind speech, and when they err there is a clearly regulated appeals process.
So I don't think it's a good idea for companies to be allowed to ban whole apps like Parler or persons like Trump, even though I principally endorse these measures. The wrong mechanisms are behind it. Instead there should be more oversight of corporations and how they make decisions that may massively affect someone's online life (which has become an almost essential part of life for most people). Mere accusations of crimes should not lead to a full-fledged ban.
I'm fine with limiting the number of people one can reach, e.g. Twitter disallowing retweets or Facebook sending posts only to closest friends. However, even these measures should be transparent, based on openly published and clear guidelines, and there must be an appeals process with several instances and humans in the loop.
Right now, the problem is that individuals and companies have no right of public online representation, and at the same time huge corporations have all the rights to cancel anyone's account without transparency or appeal. That is inherently unjust and will get worse the more important online features become for the rest of your life, e.g. what do employers think when you apply without Facebook account, and so on.