What with war in Ukraine, Ron DeSantis using migrants for a campaign PR stunt, Trump’s legal travails, and devastating hurricanes, it’s not that surprising that one of the most important legal decisions in recent memory has gone largely unnoticed. The decision by a three-judge panel for the Fifth Circuit, which I wrote about for Fast Company last week, upheld HB20, a Texas law that prohibits large social-media platforms like Twitter, YouTube, Instagram, and Facebook, from doing almost any content moderation at all, aside from banning things like porn, explicit violence or incitement to violence, and harassment.

What that means, in simple terms, is that social-media platforms cannot ban Nazis or racists or trolls or anyone else, no matter how obnoxious or objectionable the content they generate is, and no matter how unpleasant they make the user experience on these platforms.

The decision is a very bad one in legal terms, as I wrote about and as Mike Masnick explained very well in a piece for TechDirt. To begin with, it denies social-media companies their 1st Amendment right not to have publish content they don’t approve of, and indeed compels them to publish content they find objectionable. On top of this, the decision contradicts Section 230 of the Communications Decency Act, which explicitly gives Internet companies of all stripes the right to curate user-generated content, and to ban any content they “consider” “objectionable.” And Section 230 lets Internet companies do this without thereby having to assume legal responsibility for the user-generated content they leave up.

The Fifth Circuit decision gets around this problem by effectively rewriting Section 230, and claiming that it only allows Internet companies to ban certain, limited kinds of content (essentially sexual, violent, and harassing). And, audacious as that is, it’s not exactly shocking, because the thing about Section 230 is that almost everyone, on both sides of the aisle, dislikes it, and has been spending the last few years trying to figure out how to repeal it or just circumvent it.

So both Trump and Biden have called for 230’s repeal. (Biden now wants “fundamental reform.”) And plenty of people in Congress feel the same way — during the 116th Congress (which was in office from 2019–2021), more than 25 bills were introduced to amend or repeal Section 230, while more than twenty such bills have been introduced during the current congressional term.

On the face of it, Section 230 does not seem like an obviously controversial provision, since its main thrust is to say that websites and social-media platforms are not legally responsible for things that their customers do. So while the New York Times, for instance, can be sued if one of its reporters publishes something false and defamatory, Twitter cannot be sued if I use my Twitter account to make false claims about Donald Trump.

That’s important, obviously, because if Twitter were legally responsible for every potentially defamatory statement its users made, it would become far more cautious and risk averse, and would monitor and regulate user content more aggressively than they currently do. Sites like Reddit and Wikipedia couldn’t exist in their current forms without legal immunity for user-generated content.

So why do so many people dislike Section 230? Oddly, while Republicans and Democrats are often united in their hostility to it, they hate it for diametrically opposed reasons. Conservatives think it makes it too easy for social-media sites to censor right-wing thought, and argue that if social-media platforms are going to act like traditional publishers, they should be able to be sued like traditional publishers. Democrats like Joe Biden, by contrast, think Section 230 makes it too easy for social-media platforms to not regulate misinformation, defamation, hate speech, incitement to violence, and so on, since sites aren’t legally responsible for users who put up any of these.

Now, it should be said that the idea that Section 230 is to blame for either of these problems is wrong. Repealing the provision would lead sites to moderate more, not less, which is exactly what conservatives don’t want. But it wouldn’t necessarily lead to a steep reduction in the kind of stuff people like Biden are concerned about — conspiracy theorizing, hate speech, and so on — since even without Section 230, the 1st Amendment generally protects media companies from being sued over that kind of material.

More important, though, repealing Section 230 would be a terrible idea because, whatever its flaws, it captured two key truths about the Internet: user-generated content is a key part of what makes it great, but that content has to be moderated in order for sites not to become toxic cesspools. The provision created the legal conditions that would let websites and, now, social-media platforms become places where a thousand user-generated could bloom. But it also lets platforms prune, and kill weeds.

But the best environments online are not free-for-alls. They’re moderated, managed environments. And the pressure to create these environments is not primarily political — it’s market-driven, and employee-driven. YouTube would be far less popular with consumers if it allowed itself to be overrun with ISIS videos and hate speech — and they face internal pressures, too, since lots of people don’t want to work for sites that they view as facilitating misinformation and hate speech. But because they have the protections of Section 230, these sites also moderate less than they would otherwise do.

That doesn’t mean, of course, that these sites have struck the right balance between excluding Nazis and allowing less toxic, if still extreme, material online. But we should be skeptical of the idea that the government would do a better job of striking that balance and we should be equally skeptical of the idea that taking away the power to moderate (as HB20 does) would improve social media. Section 230 may be an imperfect bit of law. But it’s almost certainly better than anything that would replace it.