Section 230 & social media censorship: Too much or too little?

Some scattered thoughts on the great social media purge of 2021 and the calls to repeal Section 230.

1) What’s the counterfactual?

Let me open by saying that I do not think repealing Section 230 would be earth shattering to big tech or “end the open internet as we know it.” When people say stuff like that, they’re either being hyperbolic or assuming an unlikely, worst-case counterfactual, a bit like when people claim overturning Roe v Wade would lead to an explosion in back-alley abortions. Advocates and activists will always tend to catastrophize, but such rhetoric is uniformly unhelpful.

Decisions are made at the margin, which requires having a conception of the realistic next-most-likely alternative. In 230's case, the realistic alternative is not no legal safe harbor for user-created content whatsoever, but rather either a modified version of the law, or a series of standards TBD by the courts.

2) Do I have my facts right?

With that said, whatever your view on Section 230, you should first check to make sure you understand what the law actually says. That main sound basic, but the law is routinely misrepresented, including in reputable sources. If you intend to critique that law, it’s particularly important to digest what people like Mike Masnick or Ben Thompson have written about it, as these are among its smartest defenders.

3) Is 230 even relevant?

Repealing 230 makes little sense as a remedy for the typical problems people cite with social media. The protections it affords extend to virtually any website with user-generated content, not just the big platform companies like Twitter or Facebook. Reforming it is thus an extremely blunderbuss option, with repercussions that (while perhaps not catastrophic) are hard to predict since there’s little consensus around what should replace it, and at least as many bad ideas circulating as good. The reason it’s brought up so often has less to do with the law itself, and more to do with using the threat of 230 repeal to pressure the big platforms into adopting different (but targeted) internal policies.

4) Are the courts the right tool?

As alluded to above, repealing Section 230 without a replacement would kick the biggest issues to the courts. Umpteen lawsuits later, new standards would begin to crystalize, and life would go on. But are courts, rather than direct legislation, really the right tool for the job? Litigation is costly, and in the bigger picture, the US’s adversarial legal system is clearly a major contributing factor to our broader economic and institutional sclerosis. I for one would like to see private rights of action scaled-back, rather than expanded in scope by several orders of magnitude.

5) Is scale a pro or con?

230 repeal makes little sense from the POV of improving viewpoint diversity, content moderation, or censorship on Facebook or Twitter. It does make some sense, however, if your goal is to knee-cap the US tech sector and make achieving scale difficult, per se.

Needless to say, that’s not my goal. On the contrary, I think big is beautiful, and that being home to tech giants is a major source of US competitiveness vis a vis Europe and Asia. The large platforms also do far more to moderate content than they would in lieu of 230’s good samaritan provisions, and indeed their scale is in many ways prerequisite to their ability to moderate.

In my view, the main ills created by social media all tend to be byproducts of algorithmic amplification. What happened at the US Capitol, for instance, takes a page right out of the internet-enabled “revolt of the public” that Martin Gurri talks about. I don’t see any easy solution to this problem, but if there is one, I suspect it will actually help to have a handful of large platforms dominating the market. Big companies are easier to regulate, and as we’re seeing right now with the Trump purge, a handful of companies can also more effectively coordinate with each other.

Personally, I’d like to see Facebook and Twitter blunt the virality of not just “fake news” but anything remotely political or polarizing while, as much as possible, fragmenting users into smaller communities so our media diets become less synchronized. This would allow such sites to preserve their efficiencies of scale without stretching our social graph too thin. Reddit, for example, is basically 100% user-generated content and thus a clear beneficiary of Section 230, and yet the site as a whole does not exhibit the usual problems associated with other social media platforms. It achieves this by segregating content into a large mosaic of opt-in communities with moderation tools built-in.

6) Too much censorship, or too little?

But isn’t it bad that a small number of tech companies are purging anything Trump related? Yes, probably. It sets a bad precedent, and could easily blow back in their face. But let’s keep things in perspective: De-platforming, even when relatively coordinated, is not the same thing as Chinese-style censorship — not even close.

As it stands, Twitter (say) lacks the capacity to censor anything. The most they can do is steer, as in steering our attention. If you’re ever banned unjustly, dozens of alternative communication media exist, including encrypted options like Signal and Keybase. Indeed, the technology for censorship-resistant everything is essentially already here. There are still kinks to work out, and for now learning curves create a modest barrier to entry. Nonetheless, the BATNA to the big social media platforms has never been better.

And while one can make the argument that private companies with sufficient market power become quasi-common carriers and thus have to play by different rules, that ignores that broader context in which, from a marginalist POV, information has never been less controlled or controllable. Even when it feels like your options are limited, the supply of new social media companies is incredibly elastic. Any serious, broad-based “censorship” will simply accelerate the adoption of end-to-end encrypted, decentralized alternatives, at which point section 230 will be rendered moot.

In that light, I find myself more worried about a future with too little social media “censorship,” not too much. Permanent Arab Spring-style revolutions does not seem like a desirable equilibrium to me, even if it occasionally results in some positive social change.