Catching Up on the 11th Circuit Appeal in NetChoice v. Moody Over Florida’s Social Media Censorship Law

As you recall, earlier this year Florida passed SB 7072, a brazenly censorial #MAGA bill. The district court enjoined the law, and Florida appealed to the 11th Circuit. In my last post, I recapped Florida’s appeal brief and the supporting amicus briefs. In this post, I’ll recap highlights from the appellees’ brief, the supporting amicus briefs, and Florida’s reply brief. (At the bottom of this post is a case library that includes links to all of the filings discussed in this post). As you’ll see, this post spends most of its words critiquing the Knight First Amendment Institute brief, which claims to oppose the Florida law but unintentionally (?) ends up helping Florida.

Appellees’ Brief

Some highlights:

“The merits of this case are not close—Florida may not commandeer private parties’ speech or require them to adopt the state’s preferred editorial choices….Florida made no secret that it enacted its unprecedented law to single out and punish certain online service providers for their perceived political viewpoints.”

The First Amendment “is not limited to traditional media, but extends to any entity that selects, arranges, and disseminates content, whether third-party or self-generated, and certainly covers entities that the government seeks to regulate because of the way they exercise their editorial discretion.”

“the Act simply does not impose anything like actual common carriers’ obligations. Far from requiring nondiscrimination, it expressly favors certain content and speakers and exempts them from the policies and standards that apply to other users. Florida invokes §230 as somehow justifying treating providers as common carriers, when the whole point of that federal law is to ensure that services do not have to become a home for all content on a non-discriminatory basis. And there is no comparison to legislative designation of telephone companies as common carriers, for telephone companies do not exercise editorial discretion over the communications they carry, and no legislature ever complained that telephone companies were forwarding a leftist agenda by blocking conservative calls. The reality is that online service providers engage in the inherently expressive editorial function, which is the very antithesis of being a mere common carrier and is a form of speech protected by the First Amendment. Florida cannot overcome that reality by slapping on the “common carrier” label.”

“Florida tries to defend the “consistency” mandate as a mere regulation of conduct, not speech. That is akin to defending a censorship regime as regulating the conduct of writing or publishing. In reality, the “consistency” mandate is all about regulating the content of editorial judgments. It compels covered providers to convey content they otherwise would not, and it establishes state authorities as the ultimate arbiters of hopelessly vague “consistency.” Such a law plainly triggers (and flunks) strict scrutiny. Giving it a non-discrimination veneer does not advance the ball.”

Amicus Briefs

Cato

“editors have a First Amendment right to make case-by-case determinations as to what speech they wish to display and what speech they wish to exclude. Florida’s law would, in the name of “consistency,” take that choice away from editors and place it in the hands of the state and its judges.”

Chamber of Progress et al

“Florida officials assert that S.B. 7072 benefits its residents. Amici strongly disagree. The Act will harm Floridians by exposing them to dangerous and objectionable content that today is screened out by provider content moderation efforts…If the Act is enforced, consumers will lose the benefits of healthy, inclusive, and widely accessible online communities and services. Providers will be forced to change or significantly reduce their moderation practices to limit liability risk or change their products in ways that could transform the services they provide. As a result, consumers will be exposed to more scams, harassment, hate speech, explicit content, and misinformation, making online services more dangerous for vulnerable populations such as children, teens, and seniors, and inhospitable to frequently targeted communities. No benefit the state may assert can outweigh the countervailing public interest in protecting consumers.”

[For more on why Florida’s bill leads to terrible policy outcomes that do not achieve the state’s purported goals, see this article.]

Knight First Amendment Institute

I appreciated the Knight First Amendment Institute’s lawsuit against Trump for engaging in censorship by capriciously blocking users on Twitter. However, more often I’ve puzzled over how the Institute interprets the “First Amendment” part of their title. For example, over the past few years, they published several essays that seemingly celebrated and advocated for clipping the First Amendment’s wings (I unfavorably critiqued one of those).

Then, in this case, their brief claims to support the appellees, but it should have been styled as “in support of neither side.” This passage illustrates the brief’s duality:

Florida’s version of the First Amendment would give the government sweeping authority over the digital public sphere and impede social media companies from addressing real harms online. Plaintiffs’ theory would make it difficult or impossible for governments to enact even carefully drawn laws intended to protect the free speech, due process, and privacy rights of platforms’ users and to ensure that our system of free expression serves democracy. Neither of these theories is defensible, and the Court should reject both of them.

To lay the foundation for clipping First Amendment protections for social media, the brief attempts to distinguish social media services from newspapers. I will more carefully dissect this compare/contrast discussion in my forthcoming paper on editorial transparency (I’ll publicly post it in mid-January), so my comments here are just a bloggish preview of that critique.

  • “whereas newspapers are comprised mainly of content they themselves create or specifically solicit, most content posted on social media platforms is generated by the platforms’ users.” [Why does/should it matter how a publisher sources its content? Offline newspapers would likely deploy additional content sourcing techniques if they had greater protection from liability and printing didn’t cost money].
  • “there is an incredible disparity in scale between newspapers and social media platforms.” [To me, the “scale” of social media is a compelling reason to give it heightened, not lessened, protections from government censorship.]
  • “newspapers are coherent speech products in a way that social media platforms are not.” [The phrase “coherent speech products” always baffles me, because it’s nonsensical (speech isn’t a “product,” and assemblages of third-party speech are rarely “coherent”) and isn’t used by the Supreme Court (I believe Prof. Volokh coined it). A number of other briefs, including the RCFP brief, rightly denigrate the phrase.]
  • Riffing on the “coherent speech products” phrase, the brief makes this empirical claim without any empirical support (it just cites other law profs): “newspapers’ readers tend to attribute newspapers’ content to the newspapers’ publishers, whereas platforms’ users do not generally attribute the content on the platforms to the platforms’ owners.” [One thing we can prove: content on Google or Facebook gets a lift in credibility through the association–that’s not the same as “attribution,” but it’s a sign that consumers build trust relationships with Internet services similar to their trust in traditional publishers.]
  • “newspapers generally do not remove content once it has been published, whereas removing content after publication is a major part of social media platforms’ operations.” [Newspapers are increasingly open to treating the content they publish as mutable. For example, many newspapers will now scrub their archival online content for privacy reasons.]
  • “newspapers rely mainly on human decision-making in order to moderate and curate content, whereas social media companies increasingly rely on machine-learning algorithms that are generally opaque even to their creators.” [This is factually inaccurate, because social media services manually taxonomize and otherwise curate content in a variety of ways. Think, for example, of the manual curation Twitter does in its “What’s Happening” feature. Furthermore, if curation is taking place, why does/should it matter if it’s done manually or on an automated basis?]
  • Based on the purported opacity difference, the brief continues: “newspapers’ decisions are explainable in a way that platforms’ decisions often are not.” [To be clear, newspaper decisions are not more transparent to readers. Readers usually have no clue why newspapers choose which stories to cover and which they don’t; how headlines are crafted; how editors decide how many words to allocate to a particular story; why stories are featured at the top and other stories are given less prominence, etc., etc. To sidestep this obvious problem, the brief uses the slippery word “explainable.” But if we never get an explanation from newspaper editors about why they made their choices, why does matter? Plus, many editors make their editorial decisions quickly and intuitively, so grilling them afterwards may not yield precise answers about why they made the choices they made. If general impressionistic descriptions are good enough, we can get that level of explainability from algorithm designers and data scientists. Given how rarely we get explanations from traditional publishers, and how often those explanations would be unenlightening even if we got them, the brief’s algorithmic exceptionalism was completely unpersuasive to me.]

Note that the brief uses newspapers as its publication exemplar, presumably in response to the powerful Miami Herald v. Tornillo precedent. However, there are many other types of editorial models–bookstores, book publishers, record labels, movie studios, etc. The the First Amendment protects them all despite their wide diversity in editorial operations and decision-making. Given the existing diversity, I see social media platforms’ editorial approaches as an additional example of the diversity enabled by the broad First Amendment umbrella.

Based on the purported legally significant differences between newspapers and social media, the brief then suggests that newspapers deserve more First Amendment protection than social media:

Some regulations that would burden editorial judgment if imposed on newspapers might not burden editorial judgment if imposed on social media companies. Even regulations that would burden social media companies’ editorial judgment might not burden that judgment to the same extent as they would burden newspapers’ editorial judgment if the regulations were imposed on them. And the government may have different reasons, and perhaps stronger ones, for imposing certain kinds of regulatory burdens on social media companies. The analogy of social media companies to newspapers is helpful—but only to a point. The similarities between platforms and newspapers are important, but, in any particular context, the differences might be important, too.

So…the similarities are important, but sometimes the differences are important too. I can’t really argue with that. ¯\_(ツ)_/¯

This takes us to the brief’s grand payoff: what regulations the Knight First Amendment Institute would be fine imposing on social media, even if it might (?) resist those obligations if imposed on newspapers:

  • platform transparency. [My editorial transparency paper will explain in some detail why I think “platform transparency” obligations are unconstitutional. I don’t see how any First Amendment supporter can review the Twitter v. Paxton litigation and still think “platform transparency” won’t accelerate censorial outcomes.]
  • due process. “An overbroad conception of editorial judgment, or an insistence that editorial judgment must be categorically immunized from regulatory burden, or an unqualified endorsement of the equation between platforms and newspapers…would make it nearly impossible for governments to establish due process protections that are important to free speech online.” [🤯 Private entities don’t owe anyone “due process.” Only the government does. The Knight First Amendment Institute knows this distinction very well. The imposition of “due process” on private speech venues has deeply censorial implications.]
  • privacy. The brief talked about targeted ads and echo chambers that are impervious to counterspeech. [I didn’t follow the brief at this point. Privacy regulations don’t always conflict with the First Amendment, so there’s no need to categorically curb the First Amendment for social media to advance many privacy goals. In other circumstances, privacy regulations absolutely restrict free speech, and that’s something we need the First Amendment to protect against.]

Overall, the brief attempts to walk a fine line, hoping that the First Amendment’s speech/press freedoms will retain their integrity even if social media gets less or no protection. I don’t see how that’s possible. Our country is embroiled in culture wars where censorship is a primary goal. The culture warriors are targeting many mechanisms for independent thinking, from books on library shelves to how school teachers cover race. The regulation of social media is a canary in the coal mine. Any censorship that can be imposed against social media will become the doctrinal springboard for censorship initiatives against other free speech enterprises, including other publishers. Fortunately, as the RCFP et al brief demonstrates, many media enterprises recognize the threat posed to them by the Florida law. I’m still scratching my head trying to figure out how the Knight First Amendment Institute got so out-of-sync with other traditional media players.

If the plaintiffs file a sur-reply, I wonder if they will correct the Knight First Amendment Institute’s problematic attempt to distinguish newspapers from social media services. I could see them viewed as sufficiently unpersuasive that no correction is needed.

The Knight First Amendment Institute filed its 11th Circuit brief before the district court issued its ruling in the Texas law (NetChoice v. Paxton). The Texas ruling implicitly disagreed with the Institute’s positions, especially when the court expressly rejected Texas’ “platform transparency” obligations. I suppose we can look forward to another Institute brief in the Fifth Circuit NetChoice v. Paxton appeal; but that brief should be filed alongside the amici supporting Texas, not the amici supporting the Internet services.

For more criticism of the Knight First Amendment Institute’s positions, see Mike Masnick and Scott Greenfield.

TechFreedom

“SB 7072 is, in short, a First Amendment train wreck.” [Given how often common carriage discussions devolve into discussing trains, this is a pun–get it?!]

Final Thoughts on the Amicus Briefs

As usual, there’s a lot of overlap among the amicus briefs, which is too bad because there were plenty of areas that would have benefited from deeper dives, such as:

  • the problems with the law’s transparency and due process requirements
  • the problems with the size-based distinctions among Internet services (see this article)
  • the bad policy implications and likely counterproductive outcomes of Florida’s law (see this article)
  • a survey of the media ecosystem to show the wide range of diverse editorial practices among publishers other than Internet services

Those points get some engagement throughout the briefs, but each of them merited a standalone brief rather than the overlapping pileup on other topics.

State’s Reply Brief

“Outside of this litigation, Plaintiffs’ members disclaim responsibility for the user speech they host, and they have won more than one lawsuit when federal courts took them at their word. Yet Plaintiffs’ principal argument to this Court is that their platforms are full of the platforms’ own speech. These representations cannot both be true. Florida identified this contradiction in its opening brief, and Plaintiffs have no answer.” [There’s no need to address the purported contradiction. The Constitution prohibits the government from censorship. Florida adopted a censorship law. It can’t do that. What’s left to discuss?]

“the Court must recognize that social media platforms are unique beasts, engaging in a dizzying array of distinct types of conduct.” [As the Reno v. ACLU opinion recognized, the Internet is indeed a unique medium. That was the basis of saying the Internet deserves the HIGHEST level of First Amendment protection–and why Florida should lose.]

“All the consistency provision does is require that platforms treat their users equally. That is a regulation of conduct, not speech, which is enough to dispatch Plaintiffs’ claims of content targeting.” [Just a reminder that people hate lawyers for making arguments like this. As discussed above, every censorship law can be unpersuasively characterized as regulations of “conduct.” Imagine this defense in Miami Herald v. Tornillo: “All our law does is regulate the conduct of manufacturing physical copies of newspapers that don’t mean our censorship standards.” Arguing that the “consistency provision” is a conduct regulation, and not a speech regulation, is mockable and embarrassing.]

“Nor can Plaintiffs turn the consistency provision’s regulation of conduct into a regulation of speech by claiming that their algorithms are expressive. After all, social media platforms in other litigation have taken the position that they neither create nor develop any content through content moderation.” [The brief (deliberately?) conflates the statutory analysis of Section 230 with the Constitutional protection for free speech. This is the kind of obviously wrong legal analysis that causes students to lose points when I see it on their final exams.]

“Plaintiffs fail even to mention the Santa Clara Principles that many of Plaintiffs’ own members have endorsed…Disclosure practices that Facebook, Twitter, and YouTube support can hardly be considered so burdensome that they should be categorically enjoined in this facial pre-enforcement lawsuit.” [First, the Santa Clara Principles expressly tell regulators not to make this argument: “The Santa Clara Principles is not a template for regulation. The principles were not created for that purpose and should not be used as such. States should not transform the Santa Clara Principles directly into legal mandates.” (I know, it was fanciful thinking by the principle drafters). Second, even if Facebook, Twitter, and YouTube can afford to do something, the law applies to many entities who are not similarly positioned. Third, even if some social media platforms are doing some of the things that the Florida law requires, it doesn’t mean that meeting all of the requirements, and legally defending those efforts against enforcement actions, is not burdensome. Fourth a reminder to all advocates who simultaneously champion digital due process while fighting government regulation–those positions have an unavoidable tension that regulators are eagerly exploiting.]

Case library (see also NetChoice’s library)