Noting that these "safeguards" are not actually safeguards.

Someone can construe plenty of things as "harmful", to attempt to, or to chill risk averse platforms (with the possibility that they might).

Either way, this infringes on the First Amendment and undermines users' free expression.