Re: I'm worried about the EU forcing prior restraint and self-censorship
-Saying odious things about group A should be legal, barring incitement to violence
Except depending on what it is you say about who, you could fall foul of laws such as the Race Relations Act prohibits. There's a few such laws covering race, gender, disability, etc.
These laws are there for the benefit of everyone, not just "ethnic minorities". They're needed because there's some members of society (generally categorised as the scum of the earth) who won't conduct themselves in a civil manner, in fact rather the opposite to the detriment of others. These laws are a very good thing.
The problem for Facebook, WhatsApp, Twitter, YouTube is that their platforms make it possible to break such laws with no real prospect of being prosecuted. You can say almost anything you like on Twitter about anyone to anyone, there's little anyone can do to actually positively identify you with sufficient rigour for a prosecution to stand up.
So the scum of the earth take the opportunity to behave appallingly, and everyone is fed up with it. Throw in the other problems (terrorism, foreign political influence, etc) and it's no wonder law enforcement agencies and politicians all over the world are beginning to act.
Fundamental Problem
The fundamental problem for the social media companies is that they're founded in America, but operate globally. The companies, born into that American environment, have tried to operate globally as they can in America. America has it's laws on what speech it thinks is permissible in a settled society (it seemingly doesn't have any), but most other countries seek to be rather more proactive in stamping out uncivilised behaviour.
The homogeneity that is Facebook, Twitter, etc is fundamentally incompatible with most other countries' laws on what speech is permitted.
The EU countries and the UK are merely moving to enforce their pre-existing laws on permitted speech. This is a good thing for society as a whole. The companies will have to bend with that, but doing so means an end to their homogeneous, one-service-fits-all approach to global service provision. They're going to have to operate in different ways in different countries.
A Different Way to Operate
Basically, in Europe and the UK to guarantee avoiding carrying the can for their user's content (which is where things are heading), the companies are going to have to be able to pass the buck on to those users in a way that satisfies the legal authorities (i.e. it has to become easier to prosecute users who break the law). Realistically that means an end to anonymous, free accounts. They'd have to know the full legal ID of their users. That can most easily be established via a financial transaction. They have to stop being free-to-use, and become paid-for services.
Note that's not the same as you're real name being displayed on the site. It's simply that if the police come asking "who is that scumbag?" the companies can say with a high degree of certainty who the user really (here's their credit card number, here's the name on their account, here's the address their card is registered, etc).
Armed with a user's real ID and charging for the service, the companies can: 1) keep banned users off their sites, 2) ditch the data-grabbing and advertising analytics they currently depend on to raise revenue, 3) become much more trusted by their users and by governments, 4) pay tax.
Personally speaking I think it would be good. Would I pay £5 per year to use all of Google, with a guarantee that they don't mine my data and don't use me as a product in the advertising market? Probably yes. Would I pay for WhatsApp? Well, we used to. BBM? Used to be £5 / month in the old days. Twitter? Nope. Facebook? Hmmmmm. Skype?
And that's the problem. A way out - to become paid for - works for some companies, but will probably kill others. It would reduce the choice of platforms; one or two would probably dominate at the expense of the others. Facebook and Twitter know that their services are pretty fundamentally not worth anyone's money, so unsurprisingly they're the ones making the most noise about things.
If that's the way it goes, the companies that survive are on to a good thing. They can throw out the vast bulk of their analytics infrastructure for a start, saving an absolute ton of cash.