The phrasing used in the proposal (https://energycommerce.house.gov/newsroom/press-releases/ec-...), around causing “emotional injury”, is so vague and could be used to violate people’s first amendment rights by proxy using regulation of private tech companies. I suppose this was the next step beyond vague classifications like “hate speech”.
At some point someone is going to claim encryption algorithms are malicious since it can be used to encrypt child pornography and start suing anyone that hosts examples of encryption.
I suspect that the real thing preventing such lawsuits is that generally in the US, the contents of speech rarely create liability. For example it seems perfectly fine to have advertising filled with blatantly fraudulent statements, as long as you add a few weasel words that destroy an expectation of objectivity.