SOCIAL MEDIA OR SOCIAL UTILITY COMPANY? DISSECTING THE FALSE DICHOTOMY BETWEEN COMMON CARRIER AND EDITORIAL DISRETION ANALYSES IN SOCIAL MEDIA REGULATION*
Robert Saavedra Teuton**
This paper analyzes the circuit split between the Fifth and Eleventh Circuits. The Fifth Circuit held that social media platforms can be regulated as common carriers—and are therefore limited in their ability to indiscriminately censure their users—while the Eleventh Circuit held just the opposite in that social media platforms enjoy inherent constitutional rights to editorial discretion in who they censure. However, both courts erroneously apply a misleading analysis, a false dichotomy; social media platforms are neither complete common carriers nor enjoy complete, inherently protected constitutional rights to editorial discretion. This paper articulates why and frames the issue in a way the Supreme Court would find receptive (using originalist arguments), analyzes the circuit split, and provides additional arguments for why the Court should uphold state attempts at regulating social media platforms.
II. In the Beginning: Background & Perspective
The American experiment in democracy is grounded in the belief that each and every one of us is inherently endowed with certain “unalienable rights.” But there is a dilemma in this experiment, a paradox: to protect individual liberty, individuals had to create a government and thereby voluntarily give up some liberty. The birth of the new government, the United States, began an epic balancing act that has lasted to this day. On one end of the balance sits individual rights, and on the other sits the rules, laws, and institutions of government that were originally created to protect those individual rights. And the point at which the balancing act sits is power—a little one way and individual liberty gains weight, a little the other and the government broadens.
Recognizing the obvious importance of power in democracy, our Founding Fathers created a government that broke up power in an effort to maintain the balance. For instance, the Separation of Powers split power up amongst the executive, legislative, and judicial branches. But even within these branches, primarily the legislative, power was still something the Founders’ treated with caution. So the Founders created two legislative bodies: the House to “represent the people” and the Senate to represent “the states.”
Of course, the Founders weren’t just concerned that the government would sua sponte exert its power over individual rights on its own. The Founders were also concerned that a government could be usurped by a class of powerful elite who “are the most difficult animals to manage, of anything in the whole theory and practice of government.” Another intent behind the new government’s split-up structure was to curtail the powers of the private elite. This apprehension of private power extended to companies which, (1) “were feared in the early days of our nation,” (2) needed state approval “to do anything at all,” (3) were considered “soulless” “contemplations of the law” that “could ‘concentrate the worst urges’” of groups of men, (4) of which “only a few hundred in the entire country existed.” In fact, business corporations were not considered as useful social actors until the 1800s.
III. The Here and Now: Where We Are
“Not in their wildest dreams could anyone in the Founding generation have imagined Facebook, Twitter, YouTube or Tiktok.” Let alone social media, the Founders could not in their wildest dreams have imagined a world where corporations (1) could be created as a right, (2) enjoy private rights to political speech and assembly, just like real people, (3) and populate the United States by the millions. But add to the modern corporate form the ability to gather enough data on millions of citizens to predict their psychological profiles, the ability to shape political discourse and thus the political landscape itself in the new and improved public square of the internet, and in these ways enjoy a weighty ability to influence behavior—the Founder’s “wildest dreams” would be more of a nightmare. Indeed, social media companies are the most powerful economic, social, and political actors yet to participate in the American democratic experiment. These powerful actors have gained their power through capturing and processing information and ideas—concepts the First Amendment was designed to protect.
Because the First Amendment secures fundamental rights on which all the other rights sit through its ability to let individuals check power through sharing information, it would be a mistake to condition its protection from the most powerful institutions based on simplistic categorical distinctions like “government” vs. “private corporation.” This is especially true when there is often a revolving door between the government and private corporations, and when the government and private corporations work in tandem to exert their power. Would the Founders really have found no problem with a private person buying all the roads in New York to prevent the New York Independent Journal from distributing The Federalist Papers? Would they have found no problem with a private person buying Philadelphia’s public squares to prevent Federalist Party members from expressing their ideas?
Again, social media companies are the most powerful social actors of our time, thus it would be a mistake to allow social media companies to infringe on their billions of users’ free speech simply because they aren’t the government. But that is just what social media companies have started doing. “Deplatforming” (to deplatform) is when a social media company removes a user or the user’s content from its website. Courts thus far have held that social media companies have editorial discretion and complete immunity to deplatform whoever they want—the only limit being the contractual terms of service users must agree to in exchange for their services, and the existence of “other law.” Thus social media companies can silence the voices of anyone on the political spectrum for any reason without recourse. When the First Amendment “presupposes the right conclusions are more likely to be gathered out of a multitude of tongues, than through any kind of authoritative selection” the actions of the powerful to authoritatively select who gets to speak puts into question if the right conclusions can be reached, and the balance of power will surely shift.
IV. Social Media and the False Dichotomy
Even without appeals to historical principles, fear, or policies, the pure logic of existing law compels social media companies to be regulated as common carriers in some of their functions. No doubt, social media has done, and has the potential to do many wonderful things: it can help people connect across the globe, it can provide a space for new marketplaces and civic engagement, it can help people organize in the physical world, etc. These pro-social functions are achieved through a variety of different products and services: (1) harvesting data, (2) promoting different users and posts, (3) direct messaging between users, (4) giving users a page or feed on which to publish their own posts (including the ability to repost content published by other users), (5) advertising throughout the social media site (on user pages and any other place), and (6) allowing users to comment on each other’s content.
By dissecting the different products and services provided by social media companies, it’s easy to spot the false dichotomy. Some products and services—like advertising—should enjoy editorial discretion that many publishers enjoy because they are analogous to the kinds of services provided by traditional publishers that the law already protects. Other products and services—direct messaging between users and giving users a page or feed to publish their own information on (hosting)—are more analogous to common carrier functions. Still, other products or services are either a close call, like promoting users and posts and allowing users to comment on each other’s content, or merely enhancing other products and services. For instance, harvested user data allows social media companies to better promote other inter-site content, allows companies to advertise more effectively, or sell the data. Thus, a social media company is neither completely like a common carrier nor completely like a traditional publisher that should enjoy unfettered editorial discretion—that is the false dichotomy.
However, an unreconcilable split has emerged between two Circuits that have each bought into the false dichotomy. The Fifth Circuit, in Netchoice, L.L.C. v. Paxton (“Paxton”), ruled that social media platforms are complete common carriers that can be regulated to prevent users from being deplatformed or restricted in most other respects. Ruling in just the opposite way, the Eleventh Circuit in Netchoice, L.L.C. v. Attorney General, Florida (“Florida AG”), held that social media companies enjoy editorial discretion to deplatform whoever they want, whenever they want.
V. Common Carrier Law v. Editorial Discretion
Before analyzing the Circuit Split, more about the distinction between common carrier law and editorial discretion will put things into context. Common carrier law evolved from the implied common-interest theory of contract law in England, and was adopted in early American jurisprudence where a widely accepted definition emerged: a private business that carries goods, in public service, that holds itself out to “all persons indiscriminately” for hire as a business. Thus, common carrier law is commonly applied to services like those offered by utility companies and telegraph companies.
The federal courts have left open two possibilities for what a common carrier might mean. The first definition comes from dicta of a SCOTUS opinion, F.C.C. v. Midwest Video Corp., which reasoned that cable television networks were not common carriers because “a common carrier service in the communications context is one that makes a public offering to provide communications facilities whereby all members of the public who choose to employ such facilities may communicate or transmit intelligence of their own design and choosing.” The second possible definition of “common carrier” comes from a D.C. Circuit holding that defined a common carrier as a business that serves quasi-public interests and that “both holds itself out indiscriminately for hire” to the public “and transmits information exactly as it is given by the client.”
On the other hand, editorial discretion emerges from the First Amendment. The Court has called “editorial discretion” by many other names; it denotes the idea that people are free to exclude who and what they want, and is used to describe decisions to exclude in a variety of contexts—not just a publisher’s ability to exclude. Editorial discretion protects “expressive conduct,” and protects judgments about what a speaker does with “others’ speech—which is sometimes the [speaker’s] own expression, other times not.” To qualify as First Amendment-protected expressive conduct, the conduct must be one where (1) a speaker intends to convey a conventionally comprehensible message that (2) could be understood through common language or understandings and where (3) the speaker uses a discrete set of words, conduct, or acts.
However, a speaker’s direct speech would be protected by other First Amendment doctrines even if editorial discretion, used colloquially, might also include an editor’s judgment to put their own speech in their publications. It would be illogical for editorial discretion to extend to non-speech and non-expressive conduct because the First Amendment does not protect non-speech and non-expressive conduct at all. Thus, for the editorial discretion doctrine to protect an activity, it must first implicate expressive conduct. Furthermore, the editorial discretion doctrine is commonly used to prevent compelled speech, which is a term of art that includes compelled expressive conduct.
Miami Herald Publ’g Co. v. Tornillo is the seminal editorial discretion case protecting publishers against compelled speech. At issue there was whether a Florida statute requiring newspapers to publish verbatim responses by political candidates after the newspaper published content that criticized the candidate’s “personal character or official record” was constitutional. The majority decision, in holding the Florida statute unconstitutional on First Amendment grounds, opined that newspapers are not “passive receptacles or conduits for news, comment and advertising.” Newspapers necessarily enjoy editorial discretion because the choice of what material goes into each publication is limited by the “size and content of [each] paper.” Moreover, decisions about how the newspaper treats “public issues and public officials—whether fair or unfair—constitute exercise of editorial control and judgment.”
Consequently, for the ways in which social media platforms are common carriers, they have reduced First Amendment rights in that they cannot use editorial discretion to ban any user they want for whatever reasons. Conversely, for the ways in which social media platforms maintain unfettered editorial discretion, they cannot be common carriers because they can unrestrictedly exclude anyone for any reason, which is contrary to the common carrier definition.
V. The Facts Behind the Circuit Split
Below is an overview of the pertinent facts of each case and a brief description of the decisions from the trial courts. The next section explains and compares the conflicting circuit court rulings, the Supreme Court’s decision to grant an application to vacate the Fifth Circuit’s stay of a preliminary injunction to halt Texas’s law, and how the Court might decide on how to regulate social media companies: as common carriers or publishers that enjoy editorial discretion. Although both circuits analyze very different disclosure laws, each court deciding differently, this paper does not analyze those issues. Also, the Plaintiffs in both cases (who are the same companies) sued Texas and Florida under multiple legal theories and claims. This paper will only address the First Amendment-related claims as they pertain to a state’s ability to regulate social media companies’ ability to deplatform or otherwise censure users.
In September 2021, the Texas Governor, Greg Abbott, signed House Bill 20 (“HB 20”) into law. In enacting the bill, the Texas legislature found that everyone in Texas had a right to freely exchange, share, and receive ideas and information, and that social media platforms—especially the larger ones—are common carriers. The bill’s political and legislative history was more important to the District Court’s reasoning than the Fifth Circuit’s.
The Texas bill defines social media platforms as “a website or app: (1) with more than 50 million active users in the United States in a calendar month, (2) that is open to the public, (3) allows users to create an account, and (4) enables users to communicate with each other ‘for the primary purpose of posting information, comments, messages, or images.’” Thus the law seeks to regulate “Facebook, Instagram, Pinterest, TikTok, Twitter, Vimeo, WhatsApp, and YouTube.”
The most important parts of the bill in terms of the court holdings are as follows. First, social media platforms are prohibited from “censoring a user based on the user’s viewpoint.” And “censure” means “to block, ban, remove, deplatform, demonetize, de-boost, restrict, deny equal access or visibility to, or otherwise discriminate against expression.” Thus, social media platforms cannot eliminate users’ ability to express or receive content based on either: a) the users’ viewpoint, b) the viewpoint expressed in users’ content, or c) users’ geographic location in or within Texas.
Second, “companies like Internet services providers, email providers, and sites and apps that ‘consist primarily of news, sports, entertainment, or other information or content that is not user generated but is preselected by the provider’ and user comments are ‘incidental to’ the content,” are not covered in the regulation. Relatedly, a social media platform can exclude user created information based on content if either: (1) the content is censured based on a request from an organization to prevent child exploitation or to prevent sexual abuse survivors from harassment, or (2) if content “directly incites criminal activity or consists of specific threats of violence against” people due to “race, color, disability, religion . . . or status as a peace officer or judge.” Under this code, however, it is unclear whether social media platforms would be able to moderate content that incites violence against someone due to political views.
Third, any Texan or person doing business in Texas who “shares or receives expression in” Texas can file suit against a social media platform for improperly censoring their viewpoint. The Texas Attorney General can also sue for violation of the bill or its potential violation. In any event, whoever sues a social media platform is limited in their remedy to an injunction and attorney’s fees. Lastly, the bill contains a robust severability clause.
After HB 20 was enacted, Netchoice, L.L.C. and Computer & Communications Industry Association (“Plaintiffs”), which are trade associations representing internet and social media companies like Facebook, Twitter, and Google, sued to strike down and later to enjoin the law. In granting the Plaintiff’s injunction to stay enforcement of the law, the court found that Plaintiffs were likely to succeed on the merits of their First Amendment claims because (1) social media companies are entitled to editorial discretion, and (2) the bill requires social media companies to disseminate content in violation of their rights to editorial discretion.
b. Florida AG
The contested Florida law is much different and more complex than the relatively well thought out Texas law. As a threshold matter, the Florida law applies only to certain “social media platforms” defined as “any information service, system, Internet search engine, or access software provider” doing business in Florida as a legal entity with either a yearly gross revenue over $100 million, or 100 million or more monthly users. Moreover, the Florida law protects three classes of users—with varying levels of protections—from platform actions: (1) “journalistic enterprises” enjoy the most protection, (2) political “candidate[s]” enjoy less protection, and (3) general users, merely labeled “user” enjoy the least protection. Ironically, “journalistic enterprise” and “candidate” are not included in the broader definition of “user;” as explained below, which creates interesting statutory outcomes, especially for candidates.
Platform actions, regulated in varying degrees, include a) “censor[ing],” which means any action to regulate, restrict, or alter in any way; b) deplatform[ing],” meaning to delete or ban a user from the platform for more than 14 days; c) engaging in “post-prioritization,” which means any action to “place feature or prioritize” content “ahead of, below, or in a more or less prominent position than others in a newsfeed, a feed, a view, or in search results;” and d) “shadow ban[ning],” meaning any action by any means to “limit or eliminate the exposure of a user or content posted by a user” to other users. Importantly, the post-prioritization regulation does not include paid-for prioritization—also known as advertisements—by any user class or third party.
To oversimplify while maintaining sound logic: (1) journalistic enterprises and their content cannot be deplatformed, shadow banned, or otherwise censured unless the journalistic enterprise posts obscene content; (2) candidates and their posts, and posts about them cannot be shadow banned or post-prioritized unless posts are obscene, but advertisements are allowed and a social media platform may deplatform and censure candidates; (3) users and their content can be deplatformed, shadow banned, and censured but advertisements are allowable. Social media platforms must censor, deplatform, and shadow ban users in a consistent manner, but the statute is silent about these things for candidates. Similarly, platforms must notify a user if the user or the user’s material is shadow banned, censored, or deplatformed; and platforms must tell users the number of other users who were provided or shown users’ content—but these provisions do not apply to journalistic enterprises nor, again, to candidates. Further, platforms must categorize the post-prioritization and shadow banning algorithms applied to users, and must allow users to opt-out of these algorithms “to allow sequential or chronological posts and content.” Lastly, the law subjects platforms to statutory, actual, and punitive damages.
The District Judge, Robert L. Hinkle, correctly framed the case up front by clarifying that platforms are not merely “like any other speaker” nor “more like [a] common carrier” but that “[t]he truth is in the middle.” After piecing out the many logical ambiguities and textual contradictions in the law itself, the Judge explained that “it cannot be said that a social media platform . . . is indistinguishable for First Amendment purposes from a newspaper. But neither can it be said that a platform engages only in conduct [and] not speech.” Unlike the Eleventh Circuit Court, the District Court did not go so far as to say that platforms enjoy general editorial discretion to deplatform, ban, or prioritize anything they want. Integral to the District Court’s holding was the fact that the law forbids platforms from indiscriminately adding disclaimers or other content to users’ posts—a fact that, as explained below, likely dooms the Florida law altogether.
Consequently, the District Court granted Plaintiffs’ preliminary injunction finding they would likely succeed on the merits, and the court did so for some (but not all) of the right reasons.
VII. The Fifth v. Eleventh Circuit: Comparing the Logic and Law
Each opinion in the circuit split bought into the false dichotomy, but on opposite ends: for the Fifth Circuit, social media platforms cannot use legal editorial discretion in any of their functions and are complete common carriers; the Eleventh Circuit found just the opposite. In doing so, each circuit was forced to use overbroad language and logic. The Fifth Circuit’s reasoning about what editorial discretion is and how it interacts with common carrier law is sound, but, again, was overbroad and did not strike down enough of the Texas law. Alternatively, the Eleventh Circuit used unsound logic and misconstrued the editorial discretion doctrine but came to the correct conclusion in striking the Florida law down. Interestingly, both opinions use the same cases—except the Eleventh Circuit used some cases within its own circuit to strike down the Florida law under an additional and almost identical First Amendment theory as editorial discretion.
a. Editorial Discretion
The first—and most important—contested point in the circuit split is what editorial discretion is. Whether editorial discretion is inherently free speech is critical to the outcome of the cases because if it is inherently free speech, then any regulations limiting platforms’ ability to moderate content are per se First Amendment violations. If not, then a more subtle analysis of how the regulations regulate, and what they regulate is necessary.
The Fifth Circuit said editorial discretion protects speech and expressive conduct; so for editorial discretion to be invoked, it must first aim to protect a particular and recognizable message. In coming to this finding, the Fifth Circuit cited three cases where editorial discretion—the discretion to exclude—was invoked to protect a speaker’s message Miami Herald, PG&E, and Hurley, and distinguished them from PruneYard and Rumsfeld. But the Fifth Circuit’s main points can be fleshed out merely by distinguishing Hurley from Rumsfeld.
In Hurley, parade organizers excluded a gay and lesbian group’s float from its St. Patrick’s Day parade for religious reasons. The Supreme Court upheld the parade organizer’s right to exclude the contested float finding that (1) the parade was a form of expression because each float was its own message even if the parade’s score did not communicate a particular message; and (2) the parade organizer’s speech was intertwined with the message it was trying to communicate in the parade. The Fifth Circuit interpreted this second point as meaning that when people rely on what the speaker thinks is important in deciding what to disseminate, like a newspaper, only then does the speaker’s judgment become a message that the editorial discretion doctrine can protect. Moreover, another’s message is more likely to be intertwined with that of the host’s, and therefore become the host’s protected speech, when the vehicle for the host’s speech is necessarily limited in size (like a parade, newspaper, newsletter, broadcast, etc.). In those cases where the host’s speech vehicle is necessarily limited, forcing the host to speak another’s message necessarily limits what the host can say and by that very fact changes the host’s message.
In Rumsfeld, the Solomon Amendment forced schools to host military recruiters and let students know about military recruiting events; law schools objected because they did not agree with the military’s sexual orientation policies and did not want others to equate the military’s policies with their own. The Court held that the statute did not interfere with the law schools’ First Amendment rights because the law schools were not forced to speak even when the schools were forced to send emails and post notices on the recruiter’s behalf.  Therefore, the law schools’ voices were not impeded.
The emails and notices, while including speech-like elements, were merely “incidental to the Solomon Amendment’s regulation of conduct”—regulating school’s non-expressive conduct of hosting—unlike “[g]overnment-mandated pledges or mottos” struck down in other well-known compelled speech cases. Crucially, the Court dismissed the argument that the law schools’ attempts to censure the military’s speech were inherently free speech because “denial of access is not inherently expressive[;] such conduct would only be understood as expressive in light of the law schools’ speech explaining it.” Similarly, just because some people could equate the military’s speech with the law school’s did not mean the law schools’ own speech was necessarily infringed.
To complete the distinction, the Fifth Circuit found that the statute in Rumsfeld was unlike the violative regulations in PG&E, Miami Herald, and Hurley because in those latter cases, the government regulations unconstitutionally interfered with “the complaining speaker’s own message [that would have been] affected by the speech it was forced to accommodate.” Therefore, the Solomon Amendment was constitutional because it did not limit what law schools could say, it did not change their own speech, nor did it require them to say anything at all.
Applying these principles, the Fifth Circuit found that social media platforms’ own message aren’t intimately connected to the messages their users disseminate. Unlike newspapers or parades, platforms have “virtually unlimited space for speech,” so platforms hosting some speech doesn’t necessarily mean they must cut out some other speech. Similarly, the court found that the Texas law did not directly affect the platforms’ own speech because they were “free to say whatever they want to distance themselves from the speech they host.”
On the other hand, the Eleventh Circuit found that editorial discretion in and of itself is expressive conduct and used two independent theories to say just that (both theories say the same thing). The first theory comes from the same cases cited by the Fifth Circuit: Miami Herald, PG&E, Turner, and Hurley; while the second theory comes from its own circuit’s cases: Coral Ridge Ministries Media, which in turn borrows its key reasoning, the definition of expressive conduct, from Food Not Bombs. Even though the court claims it used separate theories, which can be described as the “editorial discretion” theory and the “inherently expressive conduct” theory, they are virtually identical. As the Eleventh Circuit explained for itself, “whether we assess social-media platforms’ content-moderation activities against the Miami Herald line of cases or against our own decisions explaining what constitutes expressive conduct, the result is the same: Social-media platforms exercise editorial judgment that is inherently expressive.”
In the first theory’s line of cases, the court cherry picked language that would support its conclusion that editorial discretion (or the ability to exclude) is inherent free speech, but without the accompanying reasoning. For example, when describing Hurley (the parade case), the court merely cited to “words equally applicable here” from that case: “the presentation of an edited compilation of speech . . . fall[s] squarely within the core of First Amendment security.” Contradicting itself, the court went on to quote Hurley further, saying that “once the parade organizer’s message was understood . . . the speech then became expressive” and thus protected. But from this language it is evident that the protections editorial discretion provides depends on the underlying conduct being expressive or not; and even by the court’s own reasoning, editorial discretion is not inherently protected.
Notwithstanding that, the court failed to consider the subtleties of the Hurley case: the fact that parades are necessarily small, the fact that parades themselves express a message or at least a theme (people go to themed parades like Christmas parades, gay rights parades, victory parades, etc.), and that these facts make parade organizers’ speech intertwined with their editorial judgment—which doesn’t always inherently express a message on its own. But the court failed to elucidate these crucial subtleties. Instead, it finished its Hurley explanation by stating that the parade organizer’s lack of a specific message was immaterial, ignoring the fact that the parade was a St. Patrick’s Day parade put on by a religious group, and held that editorial discretion is just ipso facto free speech. However, the subtilties make the difference between when parade organizers are speaking through their parade—in which case editorial discretion protects their voice—and when the parade organizers are not speaking, and therefore no voice needs to be protected.
Later in the opinion, the court claims that a reasonable observer would think a platform’s actions to deplatform a user would infer a message of disapproval. But it does not address how a user would ever know another has been deplatformed. If a user stopped seeing another user, it could be for a number of reasons (the user stopped posting information for whatever reason, the user deleted their account, the user’s account was hacked, etc.)—and platforms don’t generally tell other users when they have deplatformed someone, for what reason, or whether someone has simply deleted or frozen their own account.
Logic is lost too in the Eleventh Circuit line of cases, which it describes as the “inherently expressive conduct” line of cases. To start, the court defined expressive conduct using the Eleventh Circuit case Food Not Bombs. That case was not about a host or anyone else challenging compelled speech, but a “non-profit organization that distributed free food in a city park to communicate its view that society should end hunger and poverty by redirecting resources away from the military.” When the city tried to stop the nonprofit, the Eleventh Circuit upheld the nonprofit’s rights to distribute food under the First Amendment because “the organization’s food-sharing events would convey ‘some sort of message’ to the reasonable observer—and were therefore a form of protected expression.”
Then, the court took that definition of expressive conduct, something that conveys some sort of message, and said all editorial discretion is inherently expressive conduct. As the court explained, it had recently stopped an attempt by a Southern Poverty Law Center designated Christian hate group in Coral Ridge from compelling Amazon to use its charity platform to receive donations. The court in Coral Ridge could have stopped the compelled use of Amazon’s charity platform by finding that Amazon expressed a particular message when it distinguished between which charities it deemed truly charitable and those that it deemed were not. Instead, the court found that all private entities’ editorial discretion—or the discretion “about whether, to what extent, and in what manner” one wants to exclude information—is inherently expressive conduct because Amazon conveyed some sort of message about the organizations it wished to support. Therefore, in the Eleventh Circuit, all private decisions to exclude are expressive conduct.
b. Common Carrier
The common carrier analysis in both opinions is short and, not surprisingly, comes out on
either side of the spectrum. Both courts say that platforms either enjoy editorial discretion or are common carriers with less First Amendment rights. The Fifth Circuit found social media platforms to be common carriers because (1) social media platforms hold themselves out to the public even if they censure some obscene content; (2) social media platforms have become a key part of national commerce and daily life; and (3) there are no alternatives to the large platforms. Consequently, Texas could regulate the social media platform industry as common carriers and accordingly, could regulate platforms’ hosting functions.
Conversely, the Eleventh Circuit quickly dispelled with the idea that social media companies could be common carriers, by ruling that they aren’t because they don’t freely hold themselves out to the public, and that they don’t purport to be a passive conduit of information. The court pointed to platforms’ terms of service, and the reserved rights within those terms to curate content, to support its conclusion.
c. Intermediate and Strict Scrutiny
The Fifth Circuit found that because the Texas law, in its entirety, did not compel social media platforms to alter their own speech or force them to speak on their own behalf, the law was not subject to strict scrutiny and was constitutional. In the court’s view, at the most, the regulations could possibly warrant intermediate scrutiny; but because they were content and viewpoint neutral speech regulations, the law easily passed that level of scrutiny when those regulations were weighed against Texas’s “important [governmental] interest in protecting the widespread dissemination of information.”
The Eleventh Circuit, after finding the Florida law infringed the First Amendment in its entirety, subjected all but two of the law’s moderation restrictions to strict scrutiny. The candidate-deplatforming restriction and the requirement that platforms provide a mechanism to opt-out of post-prioritization algorithms were not content-based and were therefore only subject to intermediate scrutiny. The other candidate moderation restrictions, and all of those concerning journalistic enterprises (deplatforming, shadow-banning, etc.), were analyzed via strict scrutiny because they restrict platforms’ editorial discretion in a content-based way; the same was true for the requirement to censor consistently. In any event, the court failed to find any compelling enough governmental interest; so it found every censorship restriction unconstitutional.
d. The Supreme Court on Paxton, and a Likely Outcome
The Supreme Court granted an application to vacate stay submitted by the Plaintiffs in the Paxton case. Thus, the district court’s preliminary injunction, in favor of the plaintiffs, was given effect again after the Fifth Circuit’s stay was reversed. Justices Kagan, Alito, Thomas, and Gorsuch would have denied the application to vacate stay; and Justice Alito, joined by Justices Thomas and Gorsuch, dissented because the plaintiffs were not “substantially likely” to succeed “under existing law.” It is plausible that at least one other Justice would join the disagreeing and dissenting Justices, or at least concur with them so as to affirm Texas’s right to censure social media platform censorship in some capacity.
However, this vote would probably not come from Justice Kavanaugh unless Texas can show the social media platforms it intends to regulate are sufficiently monopolistic. While on the D.C. Circuit, then-Judge Kavanaugh opined that “the First Amendment bars the Government from restricting the editorial discretion of Internet service providers, absent a showing that an Internet service provider possesses [sufficient] market power.”
e. How the Courts Should Have Decided
I agree with the Fifth Circuit’s definitions of editorial discretion and common carrier. Similarly, I agree that platforms’ host and direct messaging functions do not implicate expressive conduct and can be regulated. Additionally, I would find that basic search functions required for users to find each other’s pages are incidental to hosting speech and can thus be regulated as well—like the emails and notices in Rumsfeld.
But I don’t necessarily agree with the court’s conclusion that demonetizing, de-boosting, or otherwise denying equal access or visibility—all things imbedded within the Texas law’s definition of “censor”—are not speech. At the very least, the court should have more deeply considered these other actions to determine whether they were speech or not. Demonetizing, de-boosting, and denying equal access (promoting) are more than mere decisions to exclude or include information; they are more than mere blind judgments about how to arrange information on a screen. They are considered decisions about what messages to promote and what should get less attention. In that way, a platform expresses a message when it promotes content because it aims to tell its users what it thinks is important and what is not; infringing on that message is altering platforms’ voices, to use the Fifth Circuit’s own definition of legal editorial discretion. Further, it is much easier to discern a message from promoted or demoted content than content that has been deleted or banished altogether. In the case of promoted content, users could more easily find a message by seeing existing content more frequently and other content less frequently. But in the case of forbidden content (oftentimes non-existent from a user’s point of view), users have to infer a message from what they do not see—which is especially hard when users do not know what information platforms consider disseminating in the first place.
Yes, I agree that these provisions would survive intermediate scrutiny considering the governmental interests the Fifth Circuit cited. But I think promotional decisions are speech and that editorial discretion may be invoked by platforms to protect that speech to prevent them from being compelled to promote content they may not want to in the absence of narrowly tailored governmental regulations that serve compelling governmental interests mandating they do otherwise. Thus, for regulations on platforms’ promotional abilities, courts should first decide whether they are content-based or content-neutral restrictions before applying the appropriate level of scrutiny. And even if strict scrutiny is triggered because a promotional regulation is content based, the regulation may still be valid based on an additional compelling governmental interest explained below.
Alternatively, I agree that almost all of the Florida law should have been struck down. The Florida law was poorly written and its provisions, often in conflict with each other, mandated strange outcomes. Many other of its provisions—like prohibiting platforms from disclaiming users’ posts—limited platforms’ own speech. Moreover, the statute was preempted by § 230 of the Communications Decency Act because it imposed monetary damage liabilities on platforms. But I do not agree with the Eleventh Circuit’s construction of editorial discretion and expressive conduct and its overgeneralization that the choice to exclude is always protected free speech. Even more importantly, I disagree with what the Eleventh Circuit did not consider to be important governmental interests; thus, I disagree with the court’s invalidation of the viewpoint- and non-content-based censorship restrictions.
Since at least Miami Herald, the Supreme Court has repeatedly held that controlling private, extremely powerful, and monopolistic news organizations that place “the power [to] . . . shape public opinion” in “a few hands,” is not a compelling government interest in light of the press’s First Amendment rights. On its face, this same logic could apply to social media platforms. But it should not because social media companies are fundamentally different; they gather vast amounts of individual user data to build psychological profiles and tailor their messages to maximize attention and behavior on an individual bases—in real time. They are not mere news organizations that disseminate information generally by geographic region.
That real time and comprehensive data collection changes the nature of an act is supported by Supreme Court precedent in other contexts. For example, in Carpenter the Court said it was impermissible for police to gather cell phone data from a cell phone provider without a warrant. Notwithstanding the long solidified third-party doctrine, theorizing that people have no privacy interest in information they willingly give up to third-parties, the Court found a privacy interest in cell-phone data such that police need a warrant to retrieve it. Central to the Court’s reasoning was that (1) cell phone data is generated automatically, (2) cell phones are indispensable to modern society, and (3) the geographic information given up through cell phone data could reveal “familial, political, professional, religious, and sexual association.”
Similarly, social media companies gather information on a real-time basis automatically and are indispensable to modern society, but reveal much more information than cell phone geographic data. The Court should find that governmental platform regulation is an extremely compelling interest. This interest is enough to overcome intermediate scrutiny of viewpoint yet content-neutral regulations; perhaps it is even enough to overcome strict scrutiny of content-based regulations in some cases if the challenged law is narrowly tailored. But the governmental interests are compelling indeed when social media platforms can regulate the new private square in ways inconceivable to the Founders.
VIII. “Awareness in Itself is Healing”
I wanted to end this paper by painting an analogy between the 2016 and 2020 elections and the 1907 economic crisis that prompted Congress to create the Federal Reserve—which was created to stabilize the economy by regulating financial institutions, setting interest rates, and controlling the money supply. As the analogy was going to go, unfettered social media combined with less meaningful election laws and corporate rights to political free speech nearly bankrupted political processes in 2016 and again in 2020. Thus, like how Congress successfully mitigated rampant economic instability that plagued the United States from the collapse of government’s first central bank in the 1800s until the Federal Reserve was created, so too can Congress mitigate political instability created by social media platforms. It’s a great analogy because banks are regulated with varying degrees of stringency depending on their size (and corresponding systemic risk to the financial system) and the functions they assume, and relatedly, a web of separate government agencies are tasked with regulating different aspects of the banking system—just like how the social media system should be regulated. I was then planning on explaining some of the ways we can revise tort, agency, and statutory laws to balance competing interests.
But ultimately, I care less (but still a lot) about that; I care more about how to heal our country. It is more than just politically bankrupt. It is morally and spiritually bankrupt. One of the foundational problems is that people feel suppressed, they feel like they don’t matter and that they have no voice. Indeed, repression, the feeling of being forgotten, and the ensuing anger that these feelings foster is precisely what Cambridge Analytica and the Trump campaign sought to invoke—they ultimately won the 2016 election. However, the Trump campaign and its backers surely weren’t the only interest groups that played puppet master by pulling on the strings of negative, addictive, and viral emotions to manipulate specific segments of the population.
As Carl Jung said, “what you resist persists.” The answer is not repression, it’s not banning people, not forcing them to create other platforms to congregate only to reenforce their own ideas. The answer is to foster debate and critical thought about facts. Although social media has recently been a major impediment to meaningful conversations about facts, it is also the most powerful and readily available solution. Platforms can replace their algorithms that promote conspiracy theories and controversy, with algorithms that promote facts and pro-social values. Even if promotional regulations infringe on platforms’ free speech, the governmental interests discussed earlier in this article should be compelling enough to uphold such regulations; this may be true even if the regulations are content based. Moreover, platforms have the legal right to fact check major voices and disclaim bad messages—they can directly use their own voice. But banning malicious voices just causes them to find other soap boxes to stand on, and their followers become concentrated into self-selected groups with no outside voices to challenge their ideas. As Justice Brandeis wisely put it, “repression breeds hate, [and] hate menaces stable government, [thus] the path of safety lies in the opportunity to discuss freely.”
- * I use the terms “social media company” and “social media platform;” these terms are synonymous. Also, utility companies are regularly regulated as common carriers. See generally Eugene Volokh, Treating Social Media Platforms Like Common Carriers?, 1 Journal of free speech 337 (2021). ↑
- ** I am a third-year law student at the University of Arizona with a special interest in the First Amendment and the values it protects. Also, I would like to extend a warm thanks to everyone on the Journal of Emerging Technology team. Any and all mistakes herein are solely my own. ↑
- The declaration of independence, para. 2 (U.S. 1776). ↑
- INS v. Chadha, 462 U.S. 919, 950 (1983); see also The Federalist No. 22, at 135 (Alexander Hamilton) (warning that if the government consolidated, “in a single body, all the most important prerogatives of sovereignty” it would “create in reality that very tyranny which the adversaries of the new Constitution either are, or affect to be, solicitous to avert.”); Joseph Story, Commentaries on the Constitution 547-58 (Thomas Cooley., 4th ed. 1873) (Public bodies, like private persons, are occasionally under the dominion of strong passions and excitements; impatient, irritable, and impetuous. A legislative body is not ordinarily apt to mistrust its own powers, and far less the temperate exercise of those powers . . . . If it feels no check but its own will, it rarely has the firmness to insist upon holding a question long enough under its own view, to see and mark it in all its bearings and relations to society.”). ↑
- The Adams-Jefferson Letters: The Complete Correspondence Between Thomas Jefferson and Abigail and John Adams, Edited by Lester J. Cappon in Chapel Hill: University of North Carolina Press for the Institute of Early American History and Culture (1959), https://press-pubs.uchicago.edu/founders/documents/v1ch15s58.html. ↑
- Susan D. Carle, Why the U.S. Founders’ Conceptions of Human Agency Matter Today: The example of Senate Malapportionment, 9 Tex. A&M L. Rev. 553, 562 (2022) (noting that John Adams considered the “division of powers within the legislative branch” as a check on the elite class). ↑
- However, the Founders also only gave white, landowning men (to make sure only certain people with [ what they considered, “the right”] skin in the game could play) the right to participate in the new government, and even then, it was still only a Republic and not a true and direct democracy. ↑
- Joseph F. Morrissey, A Contractarian Critique of Citizens United, 15 U. Pa. Const. L. 765, 781 (2013) (citing Citizens United v. Fed. Election Comm’n, 558 U.S. 310, 427 (2010) (J. Stevens, concurring in part and dissenting in part) (noting that “Thomas Jefferson famously fretted that corporations would subvert the Republic.”)). ↑
- Morrissey, supra note 6 at 781 (citing Citizens United v. Federal Election Com’n, 558 U.S. 310 (2010) (J. Stevens, dissenting). ↑
- Id. at 782 (citing Trs. of Dartmouth Coll. v. Woodward, 17 U.S. 518, 636 (1819) (J. Marshall)). ↑
- Id. at 781 (citing Citizens United, 558 U.S. (2010) (J. Stevens, dissenting). ↑
- Id.; See generally JRE Clips, Barbara Frees Says Industrial Denial Dates Back to the British Slave Trade, YouTube (Jul. 1, 2020), https://www.youtube.com/watch?v=iQAqqkUOFc8. ↑
- Citizens United, 558 U.S. at 427 (J. Stevens, dissenting). ↑
- Netchoice, LLC. v. Att’y Gen., Fla., 34 F.4th 1196, 1203 (11th Cir. 2022) [hereinafter “Florida AG”]. ↑
- See Citizens United v. Fed. Election Comm’n, 558 U.S. 310 (2010). ↑
- See generally The Social Dilemma (Netflix 2020) (interviewing former Facebook managers and executives). ↑
- See generally Christopher Wylie, Mindf*ck: Cambridge Analytica and the Plot to Break America (Random House First ed. 2019); Eugene Volokh, Treating Social Media Like Common Carriers?, 1 J. Free Speech L. 377, 380 (2021) (citing Jonathan Zittrain, Engineering an Election, 127 Harv. L. Rev. F. 335 (2014) noting that in close elections “even small interferences with various groups’ ability to affect public opinion can make a big difference in outcomes.”) ↑
- See generally The Social Dilemma (Netflix 2020) (describing how the entire business model of social media companies like Facebook is based on their ability to influence behavior, and it is because they are at least somewhat successful at influencing behavior that they are able to earn profits and attract capital )). ↑
- There are no real alternatives to the major social media platforms, and these companies bring in a huge amount of revenue and spend lobbying huge sums on lobbying. Because of how they “voice” themselves—through their users, platforms, may be able to skirt election laws altogether. See generally Wylie supra note 16; Brittain v. Twitter, 2019 WL 2423375 (N.D. Cal. 2019) (dismissing a Senatorial candidate’s suit, claiming in part that Twitter violated federal election laws, after Twitter deplatformed him because the suit sought to impose liability on Twitter in violation of section 230 of the Communication Decency Act). ↑
- Ideas and information shape our worldview, which ultimately shape our physical world; the Founders surely knew this when they added the Bill of Rights and its First Amendment to the Constitution. See., e.g. Whitney v. California, 274 U.S. 357, 375-76 (1927) (noting that the Founders “believed that the freedom to think as you will and to speak as you think” were vital to ultimately securing a stable government). This new government and the individualistic ideas it was founded on, combined with the idea that gold and vast fortunes existed in the Western frontier (the American Dream), shaped the physical landscape of the Western United States in a way that was absent in Canada—where these ideas and laws were not as prevalent. See, e.g., Charles F. Wilkinson, Crossing the Next Meridian: Land, Water, and the Future of the West (Island Press, Reprint eds. 1993). Whereas the Western United States was quickly developed, and the environmental impact changed the shaped of rivers, forests, and wetlands, Canada’s natural resources were more methodically developed, and its land was left relatively intact. Id. The idea that intangible thoughts have tangible impacts on our bodies is also a well-documented principle of medicine. See, e.g., Fabrizio Benedetti, Placebo Effects: Understanding the Mechanisms in Health and Disease (Oxford University Press, 1st ed. 2008). ↑
- See, e.g. New York Times Co. v. Sullivan, 376 U.S. 254, 269-71 (1964). ↑
- Donald Trump and his cabinet are prime examples, but so are most other administrations. Also, remember the rumors about when Mark Zuckerberg wanted to run for President before the Cambridge Analytica story came out? Although he denied the rumor, he did hire President Barack Obama’s political pollster and Hillary Clinton’s campaign consultant for (supposedly just) a philanthropic project. See Shawn M. Carter, More Signs Point to Mark Zuckerberg Possibly Running for President in 2020, CNBC (Aug. 15, 2017, 1:51 PM), https://www.cnbc.com/2017/08/15/mark-zuckerberg-could-be-running-for-president-in-2020.html. ↑
- See Ken Klippenstein and Lee Fang, Truth Cops, The Intercept (Oct. 31, 2022, 2:00 AM), https://theintercept.com/2022/10/31/social-media-disinformation-dhs/ (describing how the Department of Homeland Security has been working to influence tech companies for years, and reporting that “Twitter, Facebook, Reddit, Discord, Wikipedia, Microsoft, LinkedIn, and Verizon Media met on a monthly basis with the FBI, CISA, and other government representatives” prior to the 2020 elections). ↑
- Would it make a difference if President Thomas Jefferson had told the public square’s private owners that the Federalist Party was really a British propaganda organization in an effort to influence the square’s private owners? Compare that scenario to the one described in David Malloy, Zuckerberg Tells Rogan FBI Warning Prompted Biden Laptop Story Censorship, BBC (Aug. 26, 2022), https://www.bbc.com/news/world-us-canada-62688532 (reporting that the FBI allegedly warned Facebook that “there was a lot of Russian propaganda in 2016” and to be aware that “there’s about to be some kind of dump that’s similar to that” right before the Biden laptop story dropped). ↑
- See, e.g. Huber v. Biden, No. 21-CV-06580-EMC, 2022 WL 827248 at *1, *6-7 (N.D. Cal. Mar. 18, 2022), aff’d, No. 22-15443, 2022 WL 17818543 (9th Cir. Dec. 20, 2022) (holding that Twitter could deplatform a user for reposting an Israeli news article that claimed an experimental COVID vaccine had a high mortality rate even though the plaintiff admitted newspaper evidence that showed that Twitter was working with the Biden administration to censure users who questioned COVID vaccines—according to the court, the evidence was not enough to show a conspiracy and even if it did, Twitter still had discretion to deplatform the plaintiff for violating its terms of service). ↑
- These terms of service are, by definition, contracts of adhesion. See Contracts of Adhesion, Blacks Law Dictionary (11th ed. 2019) (defining contracts of adhesion as a “standard-form contract prepared by one party, to be signed by another party in a weaker position, usually a consumer, who adheres to the contract with little choice about the terms.”) When combined with the fact that an increasing number of services are designed to link to a social media site (Uber, Airbnb, Venmo, etc.) making these services harder to use without a social media account, and the fact that there are few social media websites to choose from, consumers are increasingly being coerced into signing these terms. ↑
- So far, “other law” only exists is in Texas, but even that law has been stayed. See Netchoice LLC v. Paxton, 142 S.Ct. 1715-16 (2022) (granting an application to vacate stay of the Texas law). ↑
- For example, another country can ask Facebook to deplatform groups that advocate for human rights and freedom of religion, and Facebook could do so with complete legal immunity. See Sikhs for Justice “SFJ”, Inc. v. Facebook, Inc., 144 F. Supp. 3d 1088, 1090, 1095-96 (N.D. Cal. 2015), aff’d 697 Fed.Appx. 526 (9th Cir. 2017) (holding that Facebook could deplatform a religious and human rights advocacy group in India after allegedly being asked by the Indian government to do so). ↑
- New York Times v. Sullivan, 376 U.S. at 254, 270 (1964) (quoting United States v. Associated Press, 52 F. Supp. 362, 372 (S.D.N.Y. 1943)). ↑
- Even ignoring social media’s political ramifications, it is still unclear if it—in its current form—is a good thing for human mental and social wellbeing. ↑
- See generally Volokh, supra note 15; Wylie supra note 15. ↑
- Cf. Sullivan, 376 U.S. 254. ↑
- This paper will, primarily, only analyze two social media functions: hosting and promoting. ↑
- 49 F.4th 439, 473 (5th Cir. 2022). ↑
- Florida AG, 34 F.4th 1196, 1200-2 (11th Cir. 2022). ↑
- Common carrier law began in mid-1600s England when “common hosts, ‘farriers,’ and common carriers” engaged in private undertakings to provide a service that touched on some kind of common interest “in a manner that complie[d] with public expectations.” Phil Nichols, Redefining ‘Common Carrier’: The Fcc’s Attempt at Deregulation By Redefinition, 18 Duke L.J. 501, 506 (1987) (quoting Lord Chief Justice Hale, A Treatise in Three Parts, in 1 Collection of Tracts Relative to the Law of England 1, 72 (F. Hargrave ed. 1787). ↑
- Nichols, supra note 34 at 508 (quoting T. Chitty & L. Temple, A Practical Treatise on the Law of Carriers of Goods and Passengers by Land, Inland Navigation, and in Ships 2, 8 (Philadelphia, T. & J.W. Johnson & Co. 1857)). ↑
- Nichols, supra note 34 at 509; see Parks v. Alta Cal. Tel. Co., 13 Cal. 422, 424-25 (Cal. 1859) (holding that a telegraph company could be a common carrier even though it was a wire service that did not purport to carry any physical goods or packages). ↑
- Nichols, supra note 34 (quoting FCC v. Midwest Video, 440 U.S. 689, 701 (1979)). ↑
- Nichols, supra note 34 (citing Nat’l Ass’n of Regul. Util. Comm’r v. FCC, 525 F.2d 630 (D.C. Cir. 1976). ↑
- Adam Candeub, Editorial Decision-Making and the First Amendment, 2 J. of Free Speech L. 1 (2022). ↑
- Id. at 1 n. 1 and accompanying text. ↑
- When discussing editorial discretion, courts often discuss the choice to exclude things or people in other contexts. See Netchoice LLC v. Paxton, 49 F.4th 439 (5th Cir. 2022) (using law schools’ and parade organizers’ decisions to exclude people from their forums as analogies to discuss editorial discretion). Therefore, this paper uses the term “editorial discretion” to discuss all kinds decisions to exclude. ↑
- Candeup, supra note 40, at 3 (explaining that editorial discretion “depends on whether editorial decisions are expressive and communicative.”). ↑
- Id. at 1 (emphasis added). ↑
- Id. at 3 (distilling a functionally equivalent test from multiple Supreme Court opinions such as Texas v. Johnson, 491 U.S. 397 (1989); Clark v. Cmty. For Creative Nonviolence, 468 U.S. 288 (1984); and Spence v. Washington, 418 U.S. 405 (1974)). ↑
- Candeup, supra note 40, at 3 (noting that the Court “has been clear that only some editorial decisions are expressive and, therefore, First Amendment protected.”) ↑
- Id. at 4. ↑
- See, e.g., Miami Herald Publ’g. Co. v. Tornillo, 418 U.S. 241, 258 (1974). ↑
- See, e.g., West Virginia Bd. of Educ. v. Barnette, 319 U.S. 624 (1943) (invalidating a compulsory flag salute law in public schools under the First Amendment). ↑
- 418 U.S. 241, 258 (1974). ↑
- Id. at 244. ↑
- Id. at 258. ↑
- Id. ↑
- Id. ↑
- Paxton, 49 F.4th at 479. ↑
- Florida AG, 34 F.4th at 1222. ↑
- Netchoice, L.L.C. v. Paxton, 573 F. Supp. 3d 1092 (W.D. Tex. 2021); Netchoice, LLC v. Moody, 546 F. Supp. 3d 1082 (2021). ↑
- Paxton, 49 F.4th 439; Florida AG, 34 F.4th 1196. ↑
- Netchoice, LLC v. Paxton, 142 S.Ct. 1715 (2022). ↑
- Compare Florida AG, 34 F.4th at 1230-31 (holding that the disclosure requirements would be constitutional but for Florida’s unduly burdensome notice and justification requirement and the fact that the penalties from noncompliance with the law could be too economically burdensome), with Paxton, 49 F.4th at 485-88 (upholding Texas’s disclosure requirements that only subject social media companies to the legal and investigative costs of noncompliance because “the First Amendment protects the Platforms from unconstitutional burdens on speech—not disclosure requirements”). ↑
- Paxton, 573 F. Supp. 3d at 1101 (reporting that Netchoice challenged Texas’s law for being void for vagueness, violating the commerce clause, violating the full faith and credit clause, violating the Fourteenth Amendment’s due process clause, being preempted by § 230 of the Communications Decency Act, violating the equal protections clause of the Fourteenth Amendment, and violating the First Amendment); Moody, 546 F. Supp. 3d at 1085 (in Florida, Netchoice alleged the Florida law violated their First Amendment rights and was unconstitutionally vague, violated its member’s equal protection rights, violated the dormant commerce clause, and was preempted by § 230 of the Communications Decency Act). ↑
- Plaintiffs also challenged the number of changes social media companies can make to their “terms of service” agreements within a specific time period and an antitrust provision in the Florida statutes. Moody, 546 F. Supp. 3d at 1088-89. ↑
- Netchoice, L.L.C. v. Paxton, 573 F. Supp. 3d 1092, 1099 (W.D. Tex. 2021). ↑
- Id. at 1107 n.3. ↑
- The bill only passed after “Governor Abbott called a special second legislative session directing the Legislature to consider” legislation “protecting social-media and email users from being censored,” and was an attempt to “allow Texans to participate on the virtual public square free from Silicon Valley censorship” or “West Coast oligarchs,” and to prevent social media from “silenc[ing] conservative viewpoints and ideas.” Paxton, 573 F. Supp. 3d at 1099, 1108, 1113, 1116 (quoting Proclamation by the Governor of the State of Texas (Aug. 5, 2021), [the Texas state] Senator Bryan Hughes (@SenBryan Hughes), Twitter (Mar. 5, 2021, 10:48 PM), Senator Bryan Hughes (@SenBryan Hughes), Twitter (Aug. 9, 2021, 4:34 PM)). These comments were cited several times throughout the District Court decision—especially in considering whether Texas had a compelling state interest to impede on Plaintiff’s First Amendment claims, but the Fifth Circuit’s opinion never mentioned these comments. Paxton, at 1116. Ostensibly, the bill was a response to Twitter banning then President Trump and over 70,000 accounts with ties to conspiracy groups related to the January 6th insurrection earlier that year. Jesus Vidales, Texas social media “censorship” law goes into effect after federal court lefts block, The Tex. Tribune (Sept. 16, 2022); see also Paxton, 573 F. Supp. 3d at 1103 (noting that the Texas Attorney General tweeted on Jan. 9 that “Twitter, Facebook, and Google” targeted conservative speech, and “vowed” to fight them as AG everything he had.”) ↑
- Paxton, 573 F. Supp. 3d at 1099 (citing Tex. Bus. & Com. Code §§ 120.001(1), 120.002(b); Tex. Civ. Prac. & Rem. Code § 143.003(c)). ↑
- Paxton, 573 F. Supp. 3d at 1099. ↑
- Id. at 1113 (quoting Tex. Civ. Prac. & Rem. Code § 143A.002 (internal quotations omitted)). ↑
- Paxton, 49 F.4th at 446 (citing Tex. Civ. Prac. & Rem. Code § 143A.002 (1)). ↑
- Id. (citing Tex. Civ. Prac. & Rem. Code § 143A.002 (a)(1)-(3)). ↑
- Paxton, 573 F. Supp. 3d at 1100 (citing Tex. Bus. & Com. Code § 120.001(1)(A)-(c)). ↑
- Id. (quoting Tex. Civ. Prac. & Rem. Code § 143A.006(a)(2)-(3)). Under this code, however, it is unclear whether social media platforms would be able to moderate content that incites violence against someone due to political views. ↑
- Id. (citing Tex. Civ. Prac. & Rem. Code § 143A.007(a),(b); §§ 143A.002(a), 143A.004(a), 143A.007). ↑
- Id. (citing Tex. Civ. Prac. & Rem. Code § 143A.008). ↑
- Id. (citing Tex. Civ. Prac. & Rem. Code §§143A.007(a),(b), 143A.008). ↑
- Tex. H.B. 20, 87th Leg., 2nd Spec. Sess., 2021 Tex. Law 14. https://capitol.texas.gov/tlodocs/872/billtext/pdf/HB00020F.pdf#navpanes=0. ↑
- Paxton, 573 F. Supp. 3d at 1101; see also Moody, 546 F. Supp. 3d at 1084. ↑
- Paxton, 573 F. Supp. 3d at 1105-09. ↑
- Id. at 1109-10. ↑
- Fla St. § 501.2041 (1)(g)(1)-(4); after onset of litigation, but before the Eleventh Circuit considered the case, Florida repealed an exception to the statutory definition of social media platform that extended to platforms that held common ownership with any large Florida theme park (Disney). Florida AG, 34 F.4th at 1205. ↑
- Fla. Stat. § 501.2041 (1)(d) (defining journalistic enterprise as an entity—not person—doing business in Florida that either: (1) “publishes in excess of 100,000 words available online with at least 50,000 paid subscribers or 100,000 monthly active users, (2) publishes 100 hours of audio or video available online with at least 100 million viewers annually, (3) operates a cable channel that provides more than 40 hours of content per week to more than 100,000 cable television subscribers, or (4) operates under a broadcast license issue by the” FCC. Under this definition many pornography websites, like Pornhub, would qualify as journalistic enterprises. ↑
- Fla. Stat. § 106.011 (3) (defining candidate as a person who seeks to qualify as a candidate for nomination, receives contributions or makes expenditures either on their own or consensually through another person, appoints a treasurer and “designates a primary depository,” and “files qualification papers and subscribes to a candidate’s oath as required by law.”) ↑
- Fla. Stat. § 501.2041 (1)(h) (defining users as people in Florida that have an account with a social media platform regardless of whether they have ever used the account). ↑
- Fla. Stat. § 501.2041 (1)(b). ↑
- Fla. Stat. § 501.2041 (1)(c). ↑
- Fla. Stat. § 501.2041 (1)(e). ↑
- Fla. Stat. § 501.2041 (1)(f). ↑
- Fla. Stat. § 501.2041 (1)(e); (2)(h); (2)(j). ↑
- Fla. Stat. § 501.2041 (2)(j) see Fla. Stat. § 847.001 (defining obscene). ↑
- See Fla. Stat. § 501.2041 (2)(h) (mandating that a candidate’s statutory protection as a candidate under the social media platform laws only lasts from “the date of qualification” to “the date of election or the date the candidate ceases to be a candidate” and that platforms “must provide each user a method by which the user may be identified as a qualified candidate”); see Fla. Stat. § 847.001 (defining obscene). ↑
- See Fla. Stat. § 501.2041 (2)(d). ↑
- Fla. Stat. § 501.2041 (2)(b). ↑
- Fla. Stat. § 501.2041 (2)(e). ↑
- Fla. Stat. § 501.2041 (2)(f). ↑
- Fla. Stat. § 501.2041 (3)(d)(6). ↑
- Moody, 546 F. Supp. 3d at 1091. ↑
- See, e.g., Moody, 546 F. Supp. 3d at 1087-88 (pointing out that it is not clear how platforms could organize content when multiple users opt to be shown content chronologically or whether “user” means one who posts or one who receives content), Id. at 1086-87 (explaining that the law contradicts itself because allowing a user to opt- out of or prioritize another user’s post by definition forces the platform to shadow ban or post-prioritize another user’s content—even if only for the requesting user). ↑
- Id. at 1093. ↑
- Compare Id. at 1092-93 (noting that “social-media providers do not use editorial judgment in quite the same way” as other private parties and publishers because the “content on their sites is, to a large extent, invisible to the provider.”) with Florida AG, 34 F.4th at 1213 (proclaiming that “[s]ocial media platforms exercise editorial judgment that is inherently expressive. When platforms choose to remove users or posts, deprioritize content in viewers’ feeds or search results, or sanction breaches of their community standards, they engage in First-Amendment-protected activity.”). ↑
- Moody, 546 F. Supp. 3d at 1093, 1096 (for example, the court rightly held that the Florida law is incompatible with § 230 of the Communications Act because the state law’s damages provisions are preempted by the federal law’s immunity from civil liability). ↑
- Again, I will not address the many theories each circuit addressed, but the Fifth Circuit started its opinion by quickly dispelling with the idea that the First Amendment overbreadth doctrine protected social media’s unfettered right to censor speech because the overbreadth doctrine seeks to protect the exchange of ideas; therefore, because the ban on censorship increased the exchange of ideas, it could not be struck down under the overbreadth doctrine. Paxton, 49 F.4th at 450-55. The court further reasoned that censorship is not the pure speech that the overbreadth doctrine seeks to protect, and instead, the court found that censorship is not speech at all. Id. (finding that “no amount of doctrinal gymnastics can turn the First Amendment’s protections for free speech into protections for free censoring.”) ↑
- Both circuits also cite the Communications Decency Act, 47 U.S.C., for the proposition that Congress’s intent behind that law supports their view. Compare Paxton, 49 F.4th at 466 (quoting § 230(c)(1) to show that Congress “instructs courts not to treat Platforms as the ‘publisher or speaker’ of user-submitted content they host.”) with Florida AG, 34 F.4th at 1221 (quoting § 223(e)(6)’s language that nothing in that “section shall be construed to treat interactive computer services as common carriers or telecommunications carriers.”). ↑
- This is a debate for scholars too. Compare Volokh, supra note 16 at 404-30 (describing editorial discretion as something that protects expressive conduct), with Adam Candeub, Editorial Decision-Making, 2 Journal of Free Speech 1, 3-9 (2022) (applying an expressive conduct test directly to editorial discretion, but noting that the Court has never equated editorial discretion with fully protected speech itself). ↑
- Paxton, 49 F.4th at 455-56, 463 (stating that “the Supreme Court’s cases do not carve out “editorial discretion” as a special category of First Amendment-protected expression. Instead, the Court considers editorial discretion as one relevant consideration when deciding whether a challenged regulation impermissibly compels or restricts protected speech.”) ↑
- Id. ↑
- Id. (citing Miami Herald v. Tornillo, 418 U.S. 241 (1974) that struck down a law requiring a newspaper to print certain political responses because the forced speech would alter the newspaper’s own speech). ↑
- Id. at 456-57 (citing PG&E v. Pub. Util. Com’n of Cal., 475 U.S. 1 (1986), a case where California impermissibly forced a utility company to carve out space in its bills to customers for its competitor to advertise its prices). ↑
- Id. at 457-58 (citing Hurley v. Irish-American Gay, Lesbian and Bisexual Group of Boston, 515 U.S. 557 (1995)). ↑
- Id. at 456-57 (citing PruneYard Shopping Ctr. V. Robbins, 447 U.S. 74 (1980), upholding a California law requiring a shopping mall to allow pamphlet distributors and signature collectors access to their premises when the law also did not infringe on the mall’s own speech). ↑
- Id. 458-59 (citing Rumsfeld v. F. for Acad. and Inst. Rts., Inc. 547 U.S. 47 (2006)). ↑
- Id. at 457-58 (citing Hurley, 515 U.S. at 560-74). ↑
- Id. ↑
- Id. at 459 (stating that “when a newspaper affirmatively chooses to publish something, it says that particular speech—at the very least—should be heard and discussed. So forcing a newspaper to run this or that column is tantamount to forcing the newspaper to speak.”). ↑
- Id. at 546; see also Volokh, supra note 16 at 404-26 (explaining that platforms are not like publishers with editorial discretion because (1) platforms have a virtually limitless medium to disseminate others’ messages, (2) people rely on newspaper publishers to filter out nonessential information to avoid information overload, and (3) newspaper consumers enjoy their publisher’s content as a coherent product). ↑
- Paxton, 49 F.4th at 458 (citing Rumsfeld, 547 U.S. at 51). ↑
- Id. (citing Rumsfeld, 547 U.S. at 61-62). ↑
- Id. ↑
- Id. (citing Rumsfeld, 547 U.S. at 61-62). See, e.g., W.Va. Bd. Of Educ. v. Barnette, 319 U.S. 624, 642 (1943) (invalidating a mandatory school pledge of allegiance requirement in public schools). ↑
- Paxton, 49 F.4th at 461 (citing Rumsfeld 547 U.S. at 66). ↑
- Id. ↑
- Id. at 459 (citing Rumsfeld, 547 U.S. at 63-65). ↑
- Id. (citing Rumsfeld, 547 U.S. at 60). ↑
- Id. at 462. ↑
- Id. ↑
- Id. Also, there was also no penalty or privilege given to anyone based on viewpoint—unlike in Miami Herald and PG&E where a duty was imposed on a newspaper to host viewpoint contrary to a previous viewpoint based on politics (Miami Herald) and where space in a billing envelope was awarded to entities based on being PG&E’s competition in the utility market (PG&E). Id.; The court went on to find two additional reasons why platforms do not exercise editorial discretion. First, platforms do not exercise editorial discretion because they do not accept “reputational [or] legal responsibility for the content [of their] edits.” Id. at 464. (distinguishing Associated Press v. NLRB, 301 U.S. 103, 127 (1937) that noted that newspapers take responsibility for deciding what is newsworthy when taking responsibility for the accuracy of the items they transmit). Second, editorial discretion involves selecting what to disseminate before content is disseminated. Id. at 464-65 (citing Ark. Educ. Television Comm’n v. Forbes, 523 U.S. 666, 674 (1998)). ↑
- Turner was not described in much depth in the Fifth Circuit case but was mentioned briefly. Paxton, 49 F.4th at 463 (citing Turner Broad. Sys., Inc. v. FCC, 512 U.S. 622 (1994)). ↑
- Florida AG, 34 F.4th at 1213. ↑
- Id. at 1214 (citing Coral Ridge Ministries v. Amazon.com, 6 F.4th 1247 (11th Cir. 2021)). ↑
- Id. (citing Fort Lauderdale Food Not Bombs v. City of Fort Lauderdale, 11 F.4th 1266 (11th Cir. 2021)). ↑
- Id. at 1212. ↑
- Id. at 1213. ↑
- See Id. at 1211-12. ↑
- Id. at 1211 (citing Hurley, 515 U.S. at 570). ↑
- Id. ↑
- Id. ↑
- The size of the speech vehicle is not really ever addressed in the Eleventh Circuit’s reasoning. In fact, the court uses Hurley and Turner (a broadcast case) to later say that it is not true that a common theme is needed to prove someone enjoys editorial discretion, Id. at 1221, but at least in those cases the size of the product being consumed (a parade and a number of broadcast stations) would still make it possible for the consumer to ascribe a common theme to the speaker’s alleged editorial discretion—an impossibility for consumers using social media as there would be a lifetime of information to scroll through to discern any kind of meaning (and even then, it is questionable whether a user could still ascertain a meaning from just looking through billions of pieces of information). ↑
- Even the Supreme Court in Hurley reasoned that although the parade’s score may not express one specific message, “each contingent’s expression in the [parade organizer’s] eyes comport[s] with the merits celebration on that day” and that forcing who to include in the parade would force the organizers to speak a message. Hurley, 515 U.S. at 574-75. ↑
- Id. at 577. ↑
- Florida AG, 34 F.4th at 1221. ↑
- Id. at 1212. ↑
- Id. ↑
- Id. ↑
- Id. (quoting Food Not Bombs, 901 F.3d at 1244-45). ↑
- Expressive conduct must be “conventionally expressive” such that a speaker intends to communicate a message that would be conventionally comprehensible. Barnes v. Glenn Theatre, Inc., 501 U.S. 560, 577 (1991) (Scalia, J., concurring). Thus, by holding that expressive conduct is something that would convey “some sort of message,” the Eleventh Circuit greatly expands what expressive conduct could be. Under the Eleventh Circuit’s definition, almost every act runs the risk of being First Amendment protected expressive conduct. See Id. at 576 (noting that “virtually every law restricts conduct, and virtually any prohibited conduct can be performed for an expressive purpose”). ↑
- Id. ↑
- Id. ↑
- That scenario would have been a clear legal use of the editorial discretion doctrine to protect a message because: Amazon would intend to communicate who it deemed charitable using a conventionally comprehensible message—allowing who it deemed truly charitable organizations access to its charity platform. But again, the Eleventh Circuit did not do that. ↑
- Id. ↑
- Id. The Eleventh Circuit went on to list various left- and right-wing social media platforms in support of the argument that all social media platforms necessarily exercise judgement. Id. at 1214. Ironically, social media platforms themselves are threatened with being deplatformed by companies like Apple and Google that run the app stores on which social media applications rely for downloads. Sarah Perez and Taylor Hatmaker, Google Blocks Truth Social from the Play Store—Will Apple be Next? TechCrunch (Sept. 1, 2022, 3:53 PM), https://techcrunch.com/2022/09/01/google-blocks-truth-social-from-the-play-store-will-apple-be-next/. Further, the platforms the court cites to make that argument that social media platforms can exclude people based on political affiliation are so small that they are hard to find on the Internet. The top three social media companies—Facebook, Google, and Twitter—dominate more than 70% of the social media market, and none of them explicitly ban anyone based on political affiliation. Top 10 Social Networking Sites by Market Share Statistics, Dreamgrow (Jan. 17, 2022), https://www.dreamgrow.com/top-10-social-networking-sites-market-share-of-visits/. Indeed, all the major platforms hold themselves out to everyone as long as a consumer agrees to the platform’s terms of service. ↑
- See Paxton, 49 F.4th at 455; Florida AG 34 F.4th at 1221. ↑
- Paxton, 49 F.4th at 474. ↑
- Id. at 472. ↑
- Id. at 476. ↑
- Id. at 479-80. ↑
- Florida AG 34 F.4th at 1221. ↑
- Id. ↑
- Paxton, 49 F.4th at 494. ↑
- Id. at 484-85. ↑
- Florida AG, 34 F.4th at 1222-26. ↑
- Id. at 1226. ↑
- Id. (explaining that, for instance, because the social media platforms could remove qualifying journalistic-enterprises’ posts for reasons other than the post’s content, that part of the law was content-based; also the restriction on promoting candidate’s posts is inherently topic based (it is at least somewhat political)). ↑
- Id. at 1222. ↑
- Id. at 1127-30. ↑
- NetChoice, LLC v. Paxton, 142 S.Ct. 1715, 1715-16 (2022). ↑
- Id. at 1716 (Alito, J., dissenting) ↑
- Volokh, supra note 16, at 422 (quoting U.S. Telecom Ass’n v. FCC, 855 F.3d 381, 418, 426-31 (D.C. Cir. 2017)). ↑
- The counter arguments to this are that (1) platforms are not screening through individual content before they filter the content (a point made by the Fifth Circuit), (2) promoting content may just be incidental to hosting speech, like the emails and notices being incidental to hosting military recruiters in Rumsfeld, and (3) it is hard to ascertain a message from what platforms promote—but if the last counterargument was true, Texas and Florida would not have felt a need to regulate this conduct in the first place. ↑
- Some jurisdictions hold that § 230 only bars liability, while other jurisdictions hold that § 230 bars suit altogether. Compare General Steel Domestic Sales v. Chumley, 840 F.3d 1178 (10th Cir. 2016) (holding that § 230 provides immunity from liability and not suit) with Hassell v. Bird, 420 P.3d 776, 789 (Cal. 2018) (holding that § 230 provides platforms immunity from suit where a plaintiff seeks an any order from the court that would otherwise fall within § 230’s parameters). ↑
- See, e.g., Miami Herald Pub. Co. v. Tornillo, 418 U.S. 241, 250-58 (1974). ↑
- See The Social Dilemma (Netflix 2020); see also Christopher Wylie, Mindf*ck: Cambridge Analytica and the Plot to Break America (First ed. 2019). ↑
- Carpenter v. United States, 138 S. Ct. 2206, 2211-20 (2018). ↑
- Id. ↑
- Id. ↑
- Id. ↑
- Id. at 2217 ↑
- See The Social Dilemma (Netflix 2020); see also Christopher Wylie, Mindf*ck: Cambridge Analytica and the Plot to Break America (First ed. 2019), ↑
- This is a quote commonly attributed to Friedrich Salomon Perls, the founder of Gestalt therapy. ↑
- See Making Sense of the Federal Reserve: History and Purpose of the Federal Reserve, Federal Reserve Bank of St. Louis, https://www.stlouisfed.org/in-plain-english/history-and-purpose-of-the-fed (last visited Apr. 26, 2023). ↑
- See generally Richard Scott Carnell, Jonathan R. Macey, Geoffrey P. Miller & Peter Conti-Brown, The Law of Financial Institutions (7th Ed. 2021). ↑
- Id. at 196-202, 223-26 (noting that banks of different sizes have different asset and capital requirements because larger banks pose larger risks to the entire economic system). ↑
- For example, national banks must obtain their charter from the Office of the Comptroller of the Currency and are henceforth supervised by that agency; bank holding companies are regulated at the holding company level by the Federal Reserve; and all federally insured depository institutions are regulated by the Federal Depository Insurance Corporation. See generally Carnell, Macey, Miller & Conti-Brown, supra note 180. Since most major banks are owned by bank holding companies, they are regulated by the three aforementioned agencies at different levels—and this is separate from the many other agencies that regulate the various other professional and financial services banks offer their customers (such as the Securities Exchange Commission, Consumer Financial Protection Bureau, State regulators, etc.). Id. Many of these different agencies are coordinated by the Financial Stability Oversight Council so that regulations are thorough, less redundant where needed, and so that agencies can better share information and analyze risks. Id. ↑
- Erica Goldberg First Amendment Contradictions and Pathologies in Discourse, 63 Ariz. L. Rev. 307, 334-38 (2022), https://arizonalawreview.org/pdf/64-2/64arizlrev307.pdf, (calling for (1) more robust education and court opinions distinguishing rational from emotional discourse, (2) harsher criminal penalties for violent actions in response to speech while rejecting the notion of “speech as violence,” and (3) making sure courts penalize “fake news” for specific and concrete harms and not general public harms). See also Toni M. Massaro & Helen Norton, Free Speech and Democracy: A Primer for Twenty-First Century Reformers, 54 U.C. Davis L. Rev. 1631, 1675-76 (2021) (reporting that regulating speakers based on identity—such as whether the speaker is a robot or human or has a foreign identity—for election purposes, could help regulate election discourse and “fake news” propaganda without favoring any specific viewpoint). ↑
- See Volokh supra note 16 at 454-460. ↑
- See Wylie supra note 169. ↑
- Id. ↑
- For example, the Russian Government funded malicious social media operations to manipulate the 2016 election, and they did so with a high level of sophistication. See, e.g. Susan Kelly, Russian Trolls Tried to Distract Voters with Music Tweets in 2016, Cornell Chronicle (Apr. 11, 2022), https://news.cornell.edu/stories/2022/04/russian-trolls-tried-distract-voters-music-tweets-2016-0. ↑
- the social dilemma (Netflix 2020). ↑
- Of course, where so much power is concentrated in an institution there is always the risk of its capture. This is why disclosure requirements are so important. One organizational structure that promotes openness and transparency is the cooperative structure. A private-governmental hybrid cooperative watchdog could provide a check on social media platforms. See generally Robert Saavedra Teuton, Developing Cooperation: Discovering Supportive Legal Frameworks and Policies for Worker Owned Cooperatives, 17, 29-30 (2018) (describing the cooperative form, its guiding values, and Italian part government-owned and part-private-owned cooperatives that provide public services). ↑
- Although extremely rare, the Court in Burson v. Freeman, 504 U.S. 191 (1992) upheld a content-based regulation prohibiting campaign materials within 100 feet of polling place entrances on election day because the restriction was narrowly tailored and served a compelling interest in preventing voter intimidation and fraud. ↑
- Whitney v. California, 274 U.S. 357, 375 (1927) (Brandeis, J., concurring). ↑