If you’re a government that wants to have substantial control over what sorts of content is available to users, then Section 230 is a problem for you. What you don’t want is a hundred million people responsible for their own posts. What you do want is a few dozen services that run the Internet and who you can say are “editorializing” and who thus face all kind of liability unless they obey specific speech codes that politicians then set. You might think that government control over platforms’ speech is prevented by the First Amendment, but that’s the insidious nature of Section 230: if you passed laws controlling how newspapers editorialize, the laws would be struck as violating the First Amendment. If instead you pass laws modifying platforms’ civil liability for users’ speech, it might have the same overall effect as the explicit laws would have on newspapers. Yet some will argue that it’s constitutional to do so.
But surely, you argue, you’re not opposed to letting firms do light moderation. They just can’t do heavy moderation, like kick off users for “political” reasons or use any algorithms to help users discover content. But nobody has a clue what “political” means here and how it differs from spam, and coming up with a definition is not the job of the US government or the courts. Similarly, does anyone seriously think that firms are going to stop using content discovery algorithms and become a pure “common carrier” of all content? Of course they won’t. They’ll just accept whatever speech codes the government develops and they’ll obey them so that they can continue to make money.
It’s depressing to see people on HN walk willingly into a speech-regulation regime for private companies, while claiming that the reason they want this is to ban algorithmic discovery. If you want to ban algorithms, just pass a law doing so and see if it’s constitutional. Instead we get this end-run where the result of these reforms is very likely to be much worse than the current situation and more critically we will still have most of the algorithms and moderation used to justify it.
Lets flip the discussion and rather than asking when section 230 should apply, lets ask when publisher ever should be held liable for published content. Platforms don't need immunity to laws regulating publisher if those laws did not exist.
If publishing laws only required that publisher prevented copyright infringement and sex trafficking, both being a requirement of existing 230, then everything could continue as before and the immunity provided by 230 would be unnecessary. The question is who if any would be harmed.
What prevents governments from using those laws to gain substantial control over what sorts of content that publisher publish? Those laws can and are occasionally used by politicians to stifle free speech.
But that's what governments do, and no idea about a nice society is going to stop that and revert to just "copyright infringement and sex trafficking".
If you’re a government that wants to have substantial control over what sorts of content is available to users, then Section 230 is a problem for you. What you don’t want is a hundred million people responsible for their own posts. What you do want is a few dozen services that run the Internet and who you can say are “editorializing” and who thus face all kind of liability unless they obey specific speech codes that politicians then set. You might think that government control over platforms’ speech is prevented by the First Amendment, but that’s the insidious nature of Section 230: if you passed laws controlling how newspapers editorialize, the laws would be struck as violating the First Amendment. If instead you pass laws modifying platforms’ civil liability for users’ speech, it might have the same overall effect as the explicit laws would have on newspapers. Yet some will argue that it’s constitutional to do so.
But surely, you argue, you’re not opposed to letting firms do light moderation. They just can’t do heavy moderation, like kick off users for “political” reasons or use any algorithms to help users discover content. But nobody has a clue what “political” means here and how it differs from spam, and coming up with a definition is not the job of the US government or the courts. Similarly, does anyone seriously think that firms are going to stop using content discovery algorithms and become a pure “common carrier” of all content? Of course they won’t. They’ll just accept whatever speech codes the government develops and they’ll obey them so that they can continue to make money.
It’s depressing to see people on HN walk willingly into a speech-regulation regime for private companies, while claiming that the reason they want this is to ban algorithmic discovery. If you want to ban algorithms, just pass a law doing so and see if it’s constitutional. Instead we get this end-run where the result of these reforms is very likely to be much worse than the current situation and more critically we will still have most of the algorithms and moderation used to justify it.