What does racism have to do with it? The crucial part of section 230 is [1]:
> No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.
Meaning, if someone makes a defamatory comment on foo.com the person who made the comment is liable not the operators of foo.com. And crucially Section 230 makes it so this holds true regardless of whether foo.com moderates comments or not. Foo.com can remove content unrelated to whatever Foo.com is about. Prior to Section 230, Foo.com had to moderate no content, otherwise it might be held liable for the content users submit to the site.
Foo.com can remove racism, but not because racism isn't protected speech. As far as government is concerned, there's no distinction between "hate speech" and any other form of speech [2]. Foo.com can remove it because it's a private company. Protections for speech for the most part only apply to the government curbing speech. Foo.com, Google, Facebook, and so on are private companies they can ban whatever arbitrary content they want - not just hate speech or racist speech. They could suddenly decide that cat photos are forbidden and ban any users and groups that post pictures containing cats. None of this is illegal, nor would it make them liable for users' content.
What's controversial here is that Google is making users curate their own comments in compliance with what Google wants. It's outsourcing its moderation to its own users. But there's really nothing new about this. Reddit has been massively successful in pushing most moderation responsibilities onto its own users. If moderators of a subreddit don't keep their subreddit clean, the whole subreddit gets shut down or has moderators removed. These news here is that Google is starting to emulate this model.
1. https://www.eff.org/issues/cda230
2. https://en.wikipedia.org/wiki/Hate_speech_in_the_United_Stat....