In the beginning there was a simple idea. If you were a platform and just acted as a conduit of information but didn't decide what got visibility - you weren't responsible for the content you transmitted - like a phone company isn't responsible for what you say on the phone. If you were a publisher - and exerted editorial control - such as deciding what gets put on the front page and what is buried 10 pages deep in the newspaper - you were responsible for the content. If you published libelous, fraudulent or other information the law held that you had decided to amplify that piece of information so were also held responsible.
In the beginning social media was a platform - you wrote something, your friends and family saw it - they really were just a conduit of information. Then social media decided they could suck up more attention by deciding what you saw - but, because its expensive to deal with libel suits they wanted to be categorized as a platform - even though they weren't. They succeeded - the good platform / publisher principle failed.
Most of the problems we are now seeing with Facebook, Amazaon, Ticktock could be solved instantly by saying - if you have a recommendation engine - you are responsible for the output of that recommendation engine. If it amplifies libelous, fraudulent or other such information - well you as the publisher are responsible.
It would mean that Ticktock, Facebook and others would drop their totally addictive design - as they would be afraid of being held responsible of the information which they are using to get users hooked. If they, by some miracle didn't drop their recommendation engine, and started acting as if they were responsible for the information their transmitted, your news feed wouldn't be filled with utter garbage. It would mean that fraudulent anti-vaxx conspiracy theories wouldn't get play on the internet. It would mean that libelous statements wouldn't get play. The list goes on and on.
Its simple, consistent with pre-existing law, and effective.