Debrief on the Taamneh v. Twitter Oral Arguments

February 25, 2023 · by Eric Goldman · in Content Regulation

My observations about the Supreme Court’s 2.5 hour long (and very tedious) oral arguments in the Taamneh v. Twitter case:

The justices struggled to define the statute’s actus reus (did Twitter take a culpable action?) and mens rea (did Twitter have a culpable mental state?) requirements, especially as applied to an automated self-service tool like Twitter.
Determining a statute’s actus reus and mens rea is standard fare for the Supreme Court. They do similar analyses all of the time. This is supposed to be one of the things they are really good at doing.
The fact that actus reus/mens rea questions reappear at the Supreme Court adds another strategic dimension to the case, because in theory justices are trying to take consistent and coherent positions on these issues from case to case.
Mens rea analyses are often irresolute. There are at least two places the analysis can go wrong. First, the court must translate the statutory scienter requirement into a sentence in English, and judges routinely disagree on the translation. Second, the court may define what evidence establishes or disproves that the actor had the requisite scienter. This is where the hypotheticals come rushing in. There is a virtually infinite number of factual variations that can arise, and it’s virtually impossible to establish a coherent rule that can elegantly accommodate this potential range, especially in the abstract.
Section 230 has largely mooted the tendentious and irresolute mens rea inquiries that vexed the justices in this oral argument, because Section 230 applies even if the defendant had a culpable mental state. Without Section 230’s application, aiding and abetting claims would have taken root a long time ago, and Internet Law would have bogged down in these irresolute questions 25+ years ago. As a result, this oral argument provides a preview of how Internet Law cases will look in a post-Section 230 world.
- [Note 1: contributory copyright infringement still turns on scienter, so we have developed some jurisprudence there (though it’s been channeled/narrowed some by the DMCA online safe harbors). It’s fair to say that the scienter elements of contributory copyright infringement has become as tortured, baroque, and irresolute as these oral arguments.]
- [Note 2: many Section 230 reform proposals have sought to reintroduce scienter into the immunity. This oral argument previews how the cases would go after such changes. Each case would turn into an irresolute slogfest that takes 2.5+ hours of Supreme Court oral advocacy. With a scienter-based exception, defendants would have a much harder time winning on a motion to dismiss, and many cases would go into costly discovery to fish for evidence showing the defendant’s culpable mental state. That’s why I view such scienter-based exceptions to Section 230 as functionally equivalent to repealing Section 230, because they would negate Section 230’s “procedural fast-lane” that’s the secret sauce of Section 230’s efficacy. The reintroduction of a scienter-based exception therefore is not a modest or minor reform to Section 230–it restructures the immunity’s architecture.]
A big point of Twitter’s defense is that it took steps to remove terrorist accounts when it learned about them. That led to an awkward question about what the result should be if Twitter didn’t try to remove terrorist accounts. As Musk is progressively turning off content moderation features and dismissing content moderation teams, this “hypothetical” isn’t as hypothetical as it might seem.
More generally, this oral argument reminded me of the pre-Section 230 advice to businesses that they should do as little content moderation as possible so that it could argue it lacked mens rea in cases like this. This was the whole point of Section 230–to discourage services from burying their head in the sand. Instead, Section 230 encourages services to do socially valuable content moderation work by not punishing them for whatever they miss.
Twitter’s lawyer, Seth Waxman, made a major concession when he said that an unsubstantiated notice from law enforcement that a service is hosting alleged terrorist activities, without any judicial oversight or other confirmation of accuracy, would confer legally chargeable knowledge. This can’t be right because law enforcement and other government actors can, and will, misuse this veto power to target political enemies. This weaponization can happen in the US (indeed, this is the basic argument over the Hunter Biden laptop story), but either way, it’s standard behavior by many foreign governments. Do we really want to give them unrestricted veto powers?
Because mens rea inquiries are often irresolute and the justices’ desire to maintain consistent positions across cases, it was impossible to get a clear picture of how the justices were likely to vote.
Waxman repeatedly declined the justices’ invitations to agree with some point or another. He clearly had talking points that had been blessed by Twitter, and he did not move off the talking points. Waxman could only concede whatever Twitter had authorized him to concede. Plus, at oral arguments, you don’t want to give an expedient concession, only to have other justices then pounce on any inconsistencies that might create.
The plaintiffs’ lawyer, Eric Schnapper, made a major concession (seemingly on the fly) in response to a question by Justice Kavanaugh. Kavanaugh pointed out that CNN had run a video interview of Osama bin Laden that surely advanced the terrorist agenda. Schnapper said that CNN would be protected from liability for running the interview due to the First Amendment. [record scratch] Say what? I emphatically agree with that, but then, why wouldn’t the same logic apply to Twitter publishing third-party content?
More generally, there was a lot of discussion about the analogy between banks and gun dealers on the one hand and Twitter on the other. These analogies were exasperating because they ignore an obvious factual distinction. Twitter is a speech venue, so any regulation of Twitter’s “outputs” necessarily affects both Twitter’s and its users’ speech. Banks and gun dealers don’t produce speech, so regulation of their outputs does not necessarily affect speech. Twitter’s lawyer didn’t hammer this point home, and I don’t know why not.
If you haven’t reviewed the complaint in this case, it may not be obvious just how unmeritorious this lawsuit is. Simply put, the plaintiffs have sued the wrong defendants. To salvage the case, the plaintiffs have kept iterating their theories and arguments over the lawsuit’s history, but this has taken them to an untenable place. Schnapper made clear that the plaintiffs were proceeding on the theory that Twitter aided and abetted by recommending terrorist content, but Waxman pointed out that in a 545 paragraph complaint, only 4 paragraphs related to recommendations. In other words, the complaint did not provide a solid enough foundation for this appeal.