Legal vs illegal cannot be enforced on a private platform because the truth procedure for "legal vs illegal" involves a judge, lawyers, often waiting for years.

What you can enforce is "so and so says it is illegal" (accurate 90% or 99% or 99.9% of the time but not 100%) or some boundary that is so far away from illegal that you never have to use the ultimate truth procedure. The same approach works against civil lawsuits, boycotts and other pressure which can be brought to bear.

I think of a certain anime image board which contains content so offensive it can't even host ads for porn that stopped taking images of cosplayers or any real life people because it eliminated moderation problems that otherwise would be difficult.

There is also spam (should spam filters for email be banned because the violate the free speech of spammers?) and other forms of disingenuous communication. When you confront a troll inevitably they will make false comparisons (e.g. banning Kiwi Farms is like banning talk to the effect that trans women could damage the legitimacy of women's sports just when people are starting to watch women's sports)

On top of that there are other parties involved. That anime site I mention above has no ads and runs at very low cost but has sustainability problems because it used to sell memberships but got cut off by payment providers. You might be happy to read something many find offensive but an advertiser might not want to be seen next to it. The platform might want to do something charitable but hosting offensive talk isn't it.

> (should spam filters for email be banned because the violate the free speech of spammers?)

I submit that spam filters should be under the sole control of their end users. If I'm using a Yahoo or Gmail account (I'm not) I should have the option to disable the spam filter entirely, or to only use personal parameters that are trained on the mail only I received, and not email should ever be summarily blackholed without letting me know in some way. If an email bounces, the sender should know. If it's just filtered, it should be in the recipient's spam folder.

> because the truth procedure for "legal vs illegal" involves a judge

This part is not correct. Private companies block what they believe to be illegal activities in their systems constantly - in order to limit the legal liability of being an accomplice to a crime. This is the case in all industries - and is standard practice from banking, to travel, to hotels, to retail... it's commonplace for companies to block services.

For spam, I would recommend that it gets a separate filter-flag allowing users to toggle it and see spam content, separately toggled from moderated content.