Thread by @PaulbernalUK

So the #OnlineSafetyBill has been delayed until the Autumn. That’s a good thing. What is also good is that at least some of the ‘free speech’ people on the right of politics have realised what a disaster the bill is to free speech. Everyone should. A thread. 1/n

A lot of focus has been on the ‘legal but harmful’ content and the ‘hurt feelings’ aspect that is covered by the #OnlineSafetyBill, but the real problems are much deeper. I’m just dealing with freedom of speech here, but there are other problems! 2/n

Firstly, don’t be under any illusions: the #OnlineSafetyBill is designed to restrict free speech. That’s its whole raison d’etre. The speech that it’s restricting is the speech deemed ‘harmful’ or speech making places ‘unsafe’ in some way. We should not hide from that. 3/n

You can argue that this speech *should* be restricted, but not that the bill isn’t designed to restrict it. Lots of speech is rightly restricted - for example child abuse imagery - the question is when is it right to restrict it, and why. 4/n

When you restrict free speech, there are a lot of questions you need to ask. Why are you restricting it? What criteria are you using? Who is deciding what those criteria are? How are decisions being made *in practice*? And, crucially, how might this go wrong? 5/n

It is easy to think of speech that *you* don’t like, and easy to decide that this speech should be restricted. Remember, though that

(a) Not everyone will agree with you; and
(b) You won’t always be the person deciding.

Imagine your opponents making the decisions. 6/n

That should make it clear. Restricting free speech is very risky. And the clever people will often *like* the restrictions you set up:

(a) They’ll work out how to evade them themselves; and
(b) They’ll work out how to trap *their* enemies with them.

This last part matters 7/n

Whenever you make an ‘anti-troll’ tool, you’ll find that the trolls use it to attack their enemies. Trolls will report their victims and try to get them banned/into trouble. And remember, many (most?) trolls don’t think they’re trolls, but think their enemies are the trolls. 8/n

Other aspects of the #OnlineSafetyBill also attack freedom of speech in ways that may not be immediately obvious. The ‘duty of care’ could be used to attack encryption and online anonymity. Fine, you might say, but *both* matter to free speech. 9/n

Free speech includes being able to communicate safely with people without the fear of being overheard or having your messages intercepted. Think of a dissident group under a repressive regime, or victims of abuse hiding from their abusers. For them, encryption matters hugely 10/n

Without it, many people will be unwilling to communicate at all. Similarly, if people are forced to use their real names they may decide it’s too dangerous to communicate publicly at all. They’re silenced: they know that if their enemies know who they are they will attack. 11/n

These are just some of the ways that the #OnlineSafetyBill *by its nature* damages free speech. This isn’t about details, or about fine-tuning it, but about the overall approach. It misunderstands free speech, and could harm it massively. We have time to think now. Use it. 12/12

P.S. There are a whole load of other ways the #OnlineSafetyBill is misjudged, but I wanted just to focus on these particular free speech ones.