Whew—after years of tortured litigation over the Texas and Florida legislatures’ efforts to impose carriage mandates on social media platforms, a wild circuit split between the Fifth and Eleventh Circuits, and the platforms’ efforts to force the Supreme Court to finally make a definitive ruling about how the First Amendment applies to the Internet, the Court has finally issued a ruling in NetChoice. And the answer is… we’re going to get another long round of litigation. So what happened?
Facial Challenges, Vague Laws, and Uncertainty All the Way Down
First off, it’s worthwhile noting the odd combination of opinions. The 5-vote majority opinion is authored by Justice Kagan, joined by Justices Sotomayor, Roberts, Kavanaugh, and Barrett (who writes a separate concurrence). Justice Jackson joins part of the Court’s opinion but concurs only in the judgement on the remainder. Justice Alito, joined by Justices Gorsuch and Thomas (who also writes separately), concurs only in the judgment.
So let’s start with the easy part: the entire Court concurs that this case was improperly teed up as a facial challenge by NetChoice, who brought the case on behalf of its tech company members. It was pretty obvious at oral argument that the Court had caught a bad vehicle, much as it had in Gonzalez v. Google, that put it at the precipice of dramatically altering Internet law without a full understanding of the implications. But unlike Gonzalez, where the Court punted on the application of Section 230 to simply conclude that the substantive law at issue didn’t apply to the platforms’ conduct, the Court didn’t have an easy way to nope out in NetChoice.
Why not? The Fifth and Eleventh Circuits’ opinions were so far apart that they left an untenable circuit split in place. The Eleventh Circuit had issued a fairly predictable and straightahead slog through the Court’s First Amendment precedent in concluding that the carriage provisions unconstitutionally abridged the platforms’ right to editorial discretion. Meanwhile, the Fifth Circuit had upheld the Texas law by imagining a parallel universe where none of the SCOTUS precedent was relevant and discerning the original public meaning of the First Amendment led to the conclusion that content moderation was an exercise of “censorship,” not “speech.” (This split is very fun for writing telecom law professors writing final exams and not so fun for the lawyers at social media platforms!)
So the Court needs a new trick to back into the bushes like Homer. And it finds that in the posture of the cases as facial challenges to the Texas and Florida laws. A facial challenge means that the law is fundamentally unconstitutional, as opposed to an as-applied challenge, which simply seeks a ruling that the law is unconstitutional as applied (or enforced) against a particular party in some particular set of circumstances.
Here, it proves easy for the Justices to variously put both parties (and the Fifth and Eleventh Circuits) on blast for the posture of the case. On the one hand, NetChoice jumped the gun by seeking a pre-enforcement challenge on the grounds that literally everything all its members ever do is protected speech. On the other hand, Texas and Florida drafted impossibly broad and vague laws, whose application and operation were totally unclear. (At oral argument, the Texas solicitor general struggled to explain which platforms the Texas law even applied to. Justice Kagan explicitly called him out for the fail—prepare carefully for oral arguments, kids!) And the Fifth and Eleventh Circuits each wnt too far in broadly concluding that the carriage provisions of the laws were (un)constitutional.
So, the Fifth and Eleventh Circuit decisions are vacated, and the cases go back for further development. But what exactly is going to happen in the next stage of litigation?
As the majority lays it out, this happens in two steps. First, the Fifth and Eleventh Circuit have to sort out exactly what activities the laws’ carriage provisions regulate and by which platforms—there might be different results with different kinds of social media platforms, messaging applications, marketplaces, and so forth. Second, the courts must figure out which of those applications violate the First Amendment (and as to which kinds of platforms). Both the majority and the dissent seem to agree that at the extremely high level of abstraction that the case was presented, there is enough murkiness here that the Court can’t issue a ruling—though the mind boggles at how complex and lengthy the analyses will need to be (the Fifth and Eleventh Circuit’s opinions collectively weigh in at around 180 pages already).
Some First Amendment Substance, Just as a ‘Lil Treat
Although both the majority and the concurrences could stop there, both contingents couldn’t help but treat themselves with some analysis of the First Amendment substance. The majority recognizes that the Fifth Circuit likely would reassert some of the bizarre excesses of its initial holding, and preemptively dismantles some of them. Justice Alito, in response, can’t help but defend the Fifth Circuit. And the additional concurrences from Justices Barrett, Jackson, and Thomas shed some additional light on how things might play out.
In particular, the majority focuses on two issues with the Fifth Circuit’s holding: whether platforms engage in expressive editorial discretion protected by the First Amendment, and the legitimacy of states’ interest in regulating that expression. Though there is more to say about these than I can possibly cover in this post, here are some preliminary thoughts.
The first issue is a threshold First Amendment inquiry: are platforms doing something expressive that amounts to speech when they engage in the kind of content moderation (or “discrimination,” if you like) regulated by the Texas and Florida laws? Again, remember that the Fifth Circuit characterizes all this stuff as non-speech “censorship,” and it’s pretty clear that no Justice is on board with going nearly that far. But the Justices seem to cover much of the continuum between the poles of “content moderation is censorship” and “content moderation is per se expressive”:
- The majority (Justice Kagan, Sotomayor, Roberts, Barrett, Kavanaugh) has the most expansive view of expressiveness by platforms. Expressly endorsing an analogical approach, the majority finds that social media platforms fit neatly into the Court’s prior jurisprudence recognizing editorial discretion (including Tornillo, PG&E, Turner, and Hurley), distinguishing PruneYard and FAIR. The majority seemingly finds the assertions of expressiveness by at least the large social media platforms to be credible, and doesn’t think it matters that platforms operating at huge scale happen to convey the vast majority of user content without moderation or that the views expressed in the content aren’t attributed to them. The majority stays focused on mainline social media platforms like YouTube and Facebook, which it seems to suspect are the primary focus of both laws, without too many digressions into the long tail of messaging, e-mail, shopping, and other sorts of platforms.
- Justice Barrett qualifies this a bit further in her separate concurrence, highlighting the complexity of assessing expressiveness across the different functions of social media platforms (e.g., messaging and search) as well as other kinds of platforms. Justice Barrett expresses particular concern about automation that goes beyond merely reflecting human editorial choices and specifically tags AI (this might be the first time “large language model” is mentioned in a SCOTUS opinion?). Justice Barrett also calls out the peril of foreign ownership in a paragraph that doesn’t specifically say “TikTok” but I have to imagine is telegraphing her thinking about the challenge to the “ban.”
- Justice Jackson, while joining some of the majority’s reasoning, likewise is not fully on board with the majority’s casual analogical approach, and struggles with defining “the right level of specificity” for assessing these questions (cf. the context-specific approach that I’ve called for).
- All the way on the other end of the spectrum, Justice Alito’s concurrence in the judgment is the most skeptical and exacting about the expressiveness of content moderation, contending that many “compilations” lack “expressive character. His opinion is winding and a little hard to follow, but it’s clear that he thinks many platforms are not engaged in expressive moderation. (He’s a little bit more on board with the majority’s analogical approach than Justice Jackson and yet reads many of the cases differently.)
The second issue—the legitimacy of Texas’s interest in regulating platforms—is part of the typical First Amendment scrutiny analysis. The Court doesn’t give us many of the details, including what level of scrutiny would apply. But the Justices do tackle Texas’s purported interest in regulating the platforms:
- The majority (not joined by Justice Jackson) concludes that Texas’s asserted interest is no more than “correct[ing] the mix of speech that the major social-media platforms present,” citing commentary from the legislature and Governor about the explicit anti-conservative-bias goals of the statute. The majority says this simply isn’t a legitimate interest, even if the laws are subject only to intermediate scrutiny. (It’ll be interesting to think through how this applies to Florida’s asserted interests, many of which were similar.) The Court goes to some lengths to distinguish Turner, where the interest was “saving the local-broadcast industry.”
- Justice Alito and Justice Thomas, by contrast, take a broader view of the states’ interests and their legitimacy. In particular, both (wrongly, in my view) call out the importance of common carriage doctrine, with Thomas expressly encouraging the courts to take it up again.
How precisely the Fifth and Eleventh Circuits will parse these issues remains to be seen. It could be that they will see the majority opinion as strongly guiding where they need to go on expressiveness and interests, or, as Justice Alito claims, that everything the majority says about expressiveness and interests is non-binding dicta.
There are also a number of other issues here, including the fate of the transparency provisions and the propriety of the Zauderer standard, and Justice Thomas’s lengthy general attack on facial challenges, that others will no doubt explore in depth. But this post is already getting too long, and we’ll have plenty of time to get into the weeds on the next season of NetChoice!
July 3 update: after sitting with this case for a couple of days, the overarching question on my mind is what exactly is going to happen on remand. The approach the Court seems to be calling for—arguably, turning the facial challenges into an exhaustive set of as-applied analyses—is going to be an absolutely enormous amount of work. I wonder if the Eleventh Circuit won’t cry uncle and just invoke Section 230 to shortcut a lot of the First Amendment hassle. This might be especially appealing given that the Moody district court already concluded that 230 preempted many of the Florida law’s carriage provisions (which as far as I can tell, the Eleventh Circuit didn’t disturb on appeal, though it did affirm that 230 didn’t preempt some of the transparency provisions). I suppose the Fifth Circuit could reverse its conclusion that NetChoice forfeited its 230 arguments and send them back to the district court to consider for the first time, though that seems less likely.