Four quadrants of Section 230ishness

Section 230 is in the news again, as the Supreme Court hears two cases challenging the liability shield. Both cases concern whether large platforms (YouTube and Twitter respectively) should be liable for a role in helping recruit or radicalize terrorists for what their algorithms recommended (YouTube) or what they insufficiently failed to suppress (Twitter).

I think it unlikely that the Supreme Court will find liability in these cases. They might well nibble at the edge of Section 230, distinguishing the present facts from facts under which they might find platform liability despite the Section 230 shield, and that might prove consequential going forward. In the Twitter case, they seem eager to consider some questions about the Anti-Terrorism Act (the act that creates the liability that Section 230 might shield). But the Court understands that serious modification of the judiciary's broad interpretation of Section 230 would dramatically upend a large swathe of the economy, with questionable democratic legitimacy.

Of course, this Court has been willing to countenace disruption and accept a loss of legitimacy where they are sufficiently ideologically motivated, as they were with abortion. But with the possible exception of Justice Thomas, I don't think there is that kind of motivation here.

It's not a balls and strikes thing. I think it's entirely reasonable, as a substantive and legal matter, to argue that contemporary algorithmic recommendation and suppression is so different in scale and kind and consequence from the kind of moderation lawmakers considered in 1996 that Section 230 protections might not apply.

But whatever the merits or demerits of that argument in theory, in practice a whole economy has been built around an aggressive interpretation of Section 230. I think it's a rotten economy, but the Court won't want to unilaterally tear it down, they'll want (probably appropriately) to punt these questions to Congress.

As readers may or may not know, I've become a Section 230 skeptic. At a theoretical level, I think Section 230 enshrines a balance of liabilities that is quite perverse. If I, in private anger, text something defamatory about a third party to a friend who then publicly tweets my text, neither of the parties meaningfully responsible for whatever injury the defamed party suffers (the forwarder or the broadcaster) can be held liable, because Section 230 shields them. Only I, who communicated under an expectation of privacy, can be. I think that's dumb.

But okay, it's an edge case. My animus towards Section 230 is about its central tendency, which has been to eliminate from the profit-and-loss statement very real diseconomies of scale associated with interpreting and curating speech well. Given the reality of network effects, eliminating these countervailing diseconomies has bequeathed to us a consolidated information sphere, dominated by massive platforms whose incentives are to promote engagement rather than high quality deliberation.

Platforms like Facebook, Twitter, YouTube, etc. are implicated in profound, undeniable harms precisely because they cannot, for example, understand whether or not a post in Myanmar is promoting extermination and genocide. But they can broadcast that post regardless, and profit from the attention it provokes.

Curating public speech well is an activity that is inherently artisanal, community and context bound. This centrifugal fact should counterbalance the yawning pull of network effects, except we've eliminated as a matter of law any reason for firms to pay attention to it. As a consequence, our public square has become Wonder Bread but worse: mass-produced and low quality, but enriched with no vitamins and minerals other than arsenic. For their trouble, the providers of this disservice have grown into prosperous and predatory monopolies, with profoundly antidemocratic power over public speech.

Section 230 seemed like a good idea at the time, but we've run the experiment. The results aren't good.

That's my view. A lot of people I usually agree with (including my past self!) vehemently support Section 230. I think I was mistaken, and they are mistaken. Yes, we want public forums on the internet, and we should craft a liability regime that renders it practical to operate one. But we want deliberation and controversial speech to inhabit a diverse array of smaller, higher quality public forums, with radically divergent "community standards". Megaplatforms may be okay for baby pictures and how-to videos. They make a terrible public square. There's no meaningful democracy with a terrible public square.

A great thing about lawsuits is they tend to go where the money is. Section 230 proponents often argue fatalistically that it's no good trying to undo the platform economy by eliminating the law they enthuse literally created it, because big platforms can afford lawyers to parry lawsuits, but small forums can't. Sure, but lots of us have blogged for decades with no liability shield, and it's fine, because defamation and other speech torts are hard to prove, and we're not so rich as to make long-shot lawsuits worthwhile.

Win a case against Facebook, thogh, and you can make millions. Yeah, they can afford lawyers. But unless they err on the side of tremendous caution, they'll need a whole lot of them.

We should strengthen anti-SLAPP protections, and define safe-harbors (perhaps a notice period before forum operators become potentially liable for third-party speech). But the blanket immunity provided by Section 230 is just, in my view, a failed experiment.

Anyway, I didn't plan to write all this. I mostly meant to post this graphic:

Positions on Section 230 mapped into four quadrants

I think it interesting that on both sides of the Section 230 discourse, there are both free-expression-focused and more censorious tendencies.

In the upper-left quadrant, we have the old-school internet utopians. They supported, and however disappointed and disillusioned now still mostly support, Section 230 as a straightforward boon to free speech. We want people to be able to express themselves easily and freely on the internet, internet forums are an important venue for that speech, but no one will operate a forum if they'll be liable for whatever crap their users say. So, it felt (and still feels to many) free-speech supportive just to eliminate liability for forum operators, regardless of how much or little those forums moderate or curate or shape content. Forums are good for speech, a liability shield enables and encourages forums, QED, it seemed at the time.

But it's worth recalling, the core of the Communications Decency Act in which Section 230 is embedded was straightfowardly to censor putatively harmful material from the internet. Section 230 was supported also by people (the upper-right quadrant) who were less concerned with promoting freewheeling expression on-line, and more interested in making it possible for forums to censor. Prior to Section 230, by moderating, a forum operator might be deemed a publisher rather than mere distributor of material, so there was a perverse (pun intended) disincentive to ban or suppress speech operators did not want on their forums.

Section 230 allowed anything without liability (pleasing the upper-left quadrant), while enabling suppression and censorship (pleasing the upper-right). Whether you were the kind of person who loved passing along transgressional memes or wanted censoriously wholesome spaces for families, as long as you were optimistic about a "marketplace of ideas" that would reward what was best, you could believe that Section 230 would serve your cause.

The bottom-right quadrant is easy to understand. A lot of people are not in favor of an anything-goes public sphere, and are happy for the law to discourage speech they think ought not go. Traditional religious conservatives have no illusions that some marketplace will elevate virtuous speech over porn. They suggest we ban the porn.

Across the political spectrum, people concerned that less regulated spaces should be safe and welcoming for people, and especially groups often subject to harassment, are under no illusion that a market will solve their problems. They argue that forums should be assiduously policed against what they deem to be harassment or hate speech, and would use the law to insist on this.

Many of those I've deemed "anti-harrassment activists" would object, I think, to the axiom that underlies my diagram, that there is a trade-off between concern about expression or safety. They would argue that in fact the two are complements, that suppressing harassment and hate-speech is in fact encouraging expression by all but the harassers and hate-speakers. Pretty much no one is opposed to "free expression", when you put it that way, but we have radically divergent views about what should count as free expression.

Regardless of the virtue of their ultimate ends, groups in the bottom right quadrant are willing to use the law to suppress outright certain categories of speech they deem objectionable. (A bit cheekily, I've placed both "woke" and "anti-woke" in the bottom-right quadrant. The anti-woke might object, they are true Free Speech Warriors! but I think it's clear that people polarized on either side of this axis are glad to use the law to block "harmful" speech from the other side. School libraries in my state are being emptied of "woke" literature, because it's "grooming" or "indoctrination".)

The quadrant I think most interesting is (perhaps unsurprisingly) the one in which I place myself. I now oppose Section 230, but perceive myself as doing so in support of the same values of free expression that once led me to celebrate that law.

There are two distinct cases for paring back Section 230 in the name of free expression. First is the case that those I've labeled "populists, conservatives, and fringes" most commonly make: Section 230 affirmatively protects platforms when they censor their views, and it should not.

If you understand the top-right quadrant, why some 90s-era social conservatives favored Section 230, then you should understand that removing the ability of forum owners to moderate without liability means classes of expression that are in practice frequently moderated away under Secion 230 would become more expressable without it.

Under the optimistic 1990s view of the internet, it was supposed to be incoherent to talk about even arguably valuable classes of expression "frequently moderated away", because the internet was going to be so open and diverse that any sort of expression anybody valued would have its home somewhere. And that is still true! You can find lots of "cancelled" perspectives at if you want to.

But that brings us to the second critique, which is mine.

In the 1990s, we didn't foresee the polarization of "freedom of reach" on the internet. On the contrary, the internet was the vehicle that would give reach to voices locked out of broadcast television and respectable newspapers. Freedom of the press, they say, is worth very little unless you can afford to operate one. The internet would give each of our "home pages" the same reach as The New York Times, if we had something of value to say. The combination of technical availability and meritocracy was going to end concerns about media oligopoly.

The consolidation of attention beneath a few large platforms has fully revived those concerns. Nazi speech would do poorly under anything I'd recognize as meritocracy, but segregating it from Twitter and onto Gab short-circuits the weighing process, and allows Nazi speakers to argue their ideas are suppressed rather than what they are, which is just dumb.

It's no good to pull an Elon Musk and just let the Nazis back, because once attention is consolidated onto megaplatforms, there is no "neutral" way for speech to be present. Megaplatform algorithms amplify and suppress, and must amplify and suppress to keep feeds appealing to the eyeballs they sell. Twitter can pretend as much as it wants that its algorithm is objective and orthogonal to the perspectives offered, but in social affairs nothing is orthogonal, choices will entrain correlations, and Nazi shit can be pretty engaging in a tabloidy way. You are either with the Nazis or against them, Elon, which will it be?

The better solution, I would argue, is for a much less polarized and hierarchical public square. We can't have a world with more than a handful of Twitters, Facebook, YouTube, TikToks. We should prefer a world with a whole lot of Gabs, except most of them would be full of decent and interesting people rather than aggrieved conspiracists. "Virality", the transformation of an individual's local expression to a signal with systematic effects should be hard, a result of consensus rather than mere "engagement". Amplification should be the result of decentralized human choices, not centralized attention-seeking algorithms.

Mastodon is imperfect in a lot of ways, but it's one "social network" that begins to approach this vision. The "local-first" ethos of the hometown Mastodon fork is an even closer fit. It augurs a civil society made up of communities that are by default local and private, from which social consensus might emerge only from meaningfully independent evaluation by diverse communities small enough to be deliberative, rather than via some signal that becomes salient and systematic among individuals by virtue of salaciousness. (Read Darius Kazemi.) We are so trained now to seek influence, each of us tiny minnows in a giant ocean hoping our next post or tweet will make us a big fish. We are rendered both pathetic and venal by this dynamic. A sane public square is a composition of diverse and sane communities, not attention-seeking atoms.

So long as the internet remains hospitable to scale and driven by monetary incentives towards engagement, we won't have diverse, sane communities. We will have giant, manipulative tabloids in the sky.

But, you might claim, these platforms "won the market test". Why should we pick winners and losers?

We did pick winners and losers, when in the 1990s we demonetized quality of curation, leaving only attention and engagement to sell. That's what Section 230 does. Absent Section 230, good moderators are financial winners, bad moderators are financial black holes.

We picked winners and losers in the 1990s when we strengthened "intellectual property law" and passed the Digital Millennium Copyright Act both of which render a mix of scale and predatory legal creativity key to market success.

There is no neutral, in law and technology. We make this stuff up. We picked winners and losers in the 1990s. We made dystopian choices. It is time to revisit them.