"Freedom of speech" does not entail "freedom of consequences" from said speech. The US also has limitations about what speech is allowed where or when (see https://en.wikipedia.org/wiki/Free_speech_zone for example). Conspiracy to commit crime is also not protected by free speech laws, many people would argue rightly so. It is not a major leap of logic to extend those anti-conspiracy laws to "speech" that incites other people to commit crimes.

A similar discussion can also be seen with Brexit in the UK: hate crimes have been steadily increasing since the Brexit campaign, and many people have been pointing at the divisive and malicious discourse surrounding that as one of the causes.

So yes, while I agree with you that speech in general should never be prohibited, and I also agree that automated takedowns of content are not desirable (mainly because once that's in place, the algorithms then become the law -- and that is not where I want to be); I also think that companies that facilitate large-scale communication between citizens can not hide behind "free speech" arguments to completely disavow any and all responsibility for consequences of speech on their platform.