Post

And you can confirm this by looking at what happened to small forums in the wake of the UK's Online Safety Act: thousands of small forums shut down because they would not be able to handle the liability

Is it that many? I know there were 5-10 examples that were regularly referred to. Very possible that someone is keeping a list…

Easier to make a list of the ones that haven't shut down.

I think 230 repeal would have an impact on large social media sites. Their algorithms would surely be considered part of their editorial voice and could be brought up in libel or defamation cases. And subject to discovery.

Such a case would fail under 1A grounds in the end. Under pre-230 law, you needed actual knowledge of content violating the law to be held liable. And no algorithm could have actual knowledge.

What about the case of an algorithm that is written to promote and amplify certain content. Perhaps even a particular lie. For example, if the algorithm specifically identified "stop the steal" lies and showed that content at a higher rate than other content. The algorithm is part of the lie.

The real issue is you NEED to understand how 230 and the First Amendment intersect. This is a key thing that many people get wrong. The underlying issue is that MOST THINGS that companies would get sued over hosting/promoting/etc are all protected by 1A. Now, that leads some to say...

"Well, then, we really don't need 230 since 1A will protect." But that's wrong, as @ericgoldman.bsky.social neatly lays out in this paper. 230 provides a procedural benefit to get rid of *frivolous* cases *that would lose* but does so much earlier. papers.ssrn.com/sol3/papers....

The *practical* difference, especially for smaller entities, is MASSIVE. Basically it will get a case tossed out for ~$100k rather than over $2mil. For a massive multi-billion company (Meta, Google, etc) with buildings full of lawyers, the difference is minimal.