Moderate the 'spam' per-user rather than globally. The main problem that I and most other users have with the anti-free-speech practices is bad categorization. Give users the tools to moderate their own experience (with defaults, so long as they can be disabled, and heck, disable the ability to change them in more repressive countries) and you basically have the problem solved.

I don't want these companies making the "disappear this" judgment calls anymore. They've shown that making those calls dispassionately and objectively isn't all that important to them.