The Future of Government Pressure on Social Media Platforms
As vast social media platforms undertake more content policing, the U.S. government has unsurprisingly tried to urge them to police things the way it prefers. This is likely to continue and, indeed, expand. What First Amendment constraints are there on such government pressure? This essay offers some tentative thoughts: 1) Some court of appeals cases have drawn lines distinguishing permissible attempts by government to persuade intermediaries to remove their users’ or business partners’ materials from impermissible government coercion. 2) The Supreme Court’s employer free speech cases may also inform our understanding of what counts as subtle coercion. 3) Courts considering other constitutional rights, especially the Fourth Amendment, have concluded that even noncoercive government persuasion may sometimes constitute impermissible evasion of the constitutional mandate. 4) A recent appellate decision (which the Supreme Court vacated on procedural grounds) suggests a potential distinction between ad hoc and systematic attempts to persuade platforms to remove content, though whether that line is ultimately either sensible or administrable is an open question.
Throughout the mid-twentieth century, many commentators sharply criticized the perceived oligarchy of mass communications. “Freedom of the press,” journalist A. J. Liebling famously said in 1960, “is guaranteed only to those who own one.”1 For a while in the early 2000s, thanks to the “cheap speech” made possible by the internet, everyone seemed to own a printing press capable of producing and distributing thousands (sometimes millions) of copies of one’s electronic leaflets.2 Many thought that the future of free speech was therefore one with broad freedom for speakers.
But now, we see it was too good to be true—for certain values of the variable “good.” It turns out that, today, we’re just borrowing printing presses: Facebook’s, X’s (formerly Twitter), YouTube’s. Even those of us who have our own blogs rely on hosting services such as WordPress, GoDaddy, and the like. And while most of the time these services are happy to let us use them, some of the time they say no. This platform interest in restricting speech has surged in the last ten years, and it seems likely to grow further.
What to do about this is one of the main free speech questions likely to occupy courts and legislatures in at least the near future. It arises in various contexts. For instance, are state laws that ban viewpoint discrimination by private platforms wise and consistent with the First Amendment, Section 230 of Title 47 of the U.S. Code (which gives internet service providers and platforms certain immunities from state regulation), and the Dormant Commerce Clause (which limits state authority to regulate interstate transactions)?3 What should we think about calls for greater “responsibility” on the part of platforms and other intermediaries?4 And, especially important, when may the government encourage or pressure social media platforms and other intermediaries to restrict speech on their property?5
We can expect greater and more organized government pressure of this sort. Some of the most important future free speech debates will be about whether courts and legislatures should step in to stop such pressure. The social media revolution has turned social media platforms into tremendously powerful political actors, capable of swaying close elections. But it has also made them relatively susceptible to pressure from foreign and domestic governments, advocacy groups, large commercial entities, and collaborations between these forces (for instance, when advocacy groups encourage both government action and advertiser boycotts).
It’s difficult for the government to control debate in thousands of newspapers or on millions of user sites, whether it tries to exert control through the threat of regulation, through the threat of congressional investigation or condemnation, or just through noncoercive attempts at persuasion. And even were it an easier task, controlling each publisher would yield only limited benefits to the government.
Some publishers may also resist regulation out of conviction—especially because it is their own speech the government is trying to control—or a business interest in continuing to cover what their competitors have stopped covering. Publishers also often have a tradition of adversarial relations with the government, so when the government asks them to remove content (or not publish it in the first place) such requests are viewed skeptically by default.
But social media platforms are more tempting targets than traditional print publishers, and they and their heirs will likely continue to be so. From the mid-2010s until today, social media entities have been persuaded to implement a range of restrictions on supposed “hate speech,” on supposed “misinformation” about medicine or elections, and even for a time on allegations that COVID-19 leaked from a Chinese government lab. Some of that persuasion (or perhaps pressure) has come from the U.S. federal government. The Supreme Court recently heard a case involving such government action, Murthy v. Missouri (2024), but dismissed it on procedural grounds.6
Such government action may have substantial costs and benefits. I don’t know with any confidence what, if anything, ought to be done about it. But I want to lay out some observations that I hope might help others to explore the matter.
Say the government urges various intermediaries—whether today’s social media platforms or, as was the case in the recent past, bookstores, billboards, or payment processors—to stop carrying certain speech. In this context, the government isn’t prosecuting them or suing them, just talking to them. Is such urging constitutional?
Generally speaking, courts of appeals have said “yes, that’s fine,” so long as the government speech merely aims to persuade the intermediaries, not to coerce by threat of prosecution, lawsuits, denial of benefits, or various other forms of retaliation. Here are some leading appellate cases so holding, which are both useful indicators of how some lower courts view such state intervention and interesting test cases for thinking about how things ought to be.
First, in 1980, a New York City official sent a letter urging department stores not to carry “a board game titled ‘Public Assistance—Why Bother Working for a Living.’” The letter said the game “does a grave injustice to taxpayers and welfare clients alike,” and closes with, “Your cooperation in keeping this game off the shelves of your stores would be a genuine public service.” Not unconstitutional, said the Second Circuit in Hammerhead Enterprises, Inc. v. Brezenoff (1983):
[T]he record indicates that Brezenoff’s request to New York department stores to refrain from carrying Public Assistance was nothing more than a well-reasoned and sincere entreaty in support of his own political perspective. . . . Where comments of a government official can reasonably be interpreted as intimating that some form of punishment or adverse regulatory action will follow the failure to accede to the official’s request, a valid claim can be stated. . . . [But] appellants cannot establish that this case involves either of these troubling situations.7
Note, though, that Brezenoff was the administrator of New York City’s Human Resources Administration, with no enforcement authority against the department stores. How might the matter have looked had he been the sheriff or the head of some civil enforcement agency?
Not long after, the U.S. Attorney General’s Commission on Pornography sent letters to various corporations (such as 7-Eleven) urging them not to sell pornographic magazines:
The Attorney General’s Commission on Pornography has held six hearings across the United States during the past seven months on issues related to pornography. During the hearing in Los Angeles, in October 1985, the Commission received testimony alleging that your company is involved in the sale or distribution of pornography. The Commission has determined that it would be appropriate to allow your company an opportunity to respond to the allegations prior to drafting its final report section on identified distributors.
You will find a copy of the relevant testimony enclosed herewith. Please review the allegations and advise the Commission on or before March 3, 1986, if you disagree with the statements enclosed. Failure to respond will necessarily be accepted as an indication of no objection.
Please call Ms. Genny McSweeney, Attorney, at (202) 724-7837 if you have any questions. Thank you for your assistance.
Not unconstitutional, said the D.C. Circuit in Penthouse International, Ltd. v. Meese (1991):
[T]he Advisory Commission had no . . . tie to prosecutorial power nor authority to censor publications. The letter it sent contained no threat to prosecute, nor intimation of intent to proscribe the distribution of the publications. . . .
We do not see why government officials may not vigorously criticize a publication for any reason they wish. As part of the duties of their office, these officials surely must be expected to be free to speak out to criticize practices, even in a condemnatory fashion, that they might not have the statutory or even constitutional authority to regulate. If the First Amendment were thought to be violated any time a private citizen’s speech or writings were criticized by a government official, those officials might be virtually immobilized.8
Third, in the late 1990s, a New York state legislator and a New York congressman accused X-Men Security—a security organization connected to the Nation of Islam—of various conspiracies, “asked government agencies to conduct investigations into its operations, questioned X-Men’s eligibility for an award of a contract supported by public funds, and advocated that X-Men not be retained.” X-Men lost certain security contracts as a result. Also not unconstitutional, ruled the Second Circuit in X-Men Security, Inc. v. Pataki (1999):
[J]ust as the First Amendment protects a legislator’s right to communicate with administrative officials to provide assistance in securing a publicly funded contract, so too does it protect the legislator’s right to state publicly his criticism of the granting of such a contract to a given entity and to urge to the administrators that such an award would contravene public policy. We see no basis on which X-Men could properly be found to have a constitutional right to prevent the legislators from exercising their own rights to speak.9
This, though, is not a uniform view. As will be noted below, the Fifth Circuit in Missouri v. Biden (2023) held that sometimes the government may violate the First Amendment by “substantially encourag[ing]” certain private parties to restrict speech, even in the absence of coercion.
On the other hand, where courts find that government speech implicitly threatened retaliation, rather than simply exhorting or encouraging third parties to block speech, they have generally found the government’s speech to be unconstitutional. The long-standing Supreme Court precedent addressing that issue is Bantam Books, Inc. v. Sullivan (1963), in which a state commission threatened to prosecute stores that sold books it deemed pornographic, including books that were protected by the First Amendment.10 Likewise, in National Rifle Association v. Vullo (2024), the Court held that the NRA could sue New York financial regulators for allegedly coercing banks and insurance companies “to cut their ties with the NRA in order to stifle the NRA’s gun-promotion advocacy.”11 And lower court cases have similarly found that there could be impermissible coercion even absent express threat of prosecution or regulatory action. Here are four such instances:
First, the mayor and a trustee of a New York town sent a letter to a newspaper demanding to learn more about who was involved in an advertisement that criticized local officials. Potentially unconstitutional, the Second Circuit held in Rattner v. Netburn (1991). Rattner, a businessman in the Village of Pleasantville, took out the critical ad in the Pleasantville Gazette, which was published by the Pleasantville Chamber of Commerce. Netburn, an elected member of the Village Board of Trustees, responded by writing a letter to the Chamber condemning the ad and asking questions about it. That was potentially an unconstitutional threat, the court held:
[The Netburn] letter stated that the recent Gazette “raises significant questions and concerns about the objectivity and trust which we are looking for from our business friends,” and it asked “[w]ho wrote” the questions and requested “a list of those members who supported the inclusion of this ‘article.’” Further, the record includes evidence that, when questioned about the letter, Netburn also stated that he had made a list of the local businesses at which he regularly shopped. . . .
[And] a threat was perceived and its impact was demonstrable. Several Chamber directors testified at their depositions that they viewed the letter as reminiscent of McCarthyism, threatening them with boycott or discriminatory enforcement of Village regulations if they permitted the publication of additional statements by Rattner; the Chamber member who had been “in charge of” the Gazette testified that following receipt of the Netburn letter, he had actually lost business and had been harassed by the Village.
Further, the Netburn letter caused the Chamber to cease publication of the Gazette; and it advised Rattner of this decision while concealing from him the fact that another issue would be forthcoming, in order to avoid having to publish in that issue material for which he had already paid. Thus, the fact that Netburn’s letter and statement “were not followed up with unannounced visits by police personnel” should hardly have been deemed dispositive since the Chamber immediately capitulated to what may reasonably be viewed as an implicit threat.12
Second, the president of the Borough of Staten Island sent a letter to a billboard company urging it to take down an antihomosexuality billboard. The letter closed with:
Both you and the sponsor of this message should be aware that many members of the Staten Island community, myself included, find this message unnecessarily confrontational and offensive. As Borough President of Staten Island, I want to inform you that this message conveys an atmosphere of intolerance which is not welcome in our Borough.
P.N.E. Media owns a number of billboards on Staten Island and derives substantial economic benefits from them. I call on you as a responsible member of the business community to please contact Daniel L. Master, my legal counsel and Chair of my Anti-Bias Task Force . . . to discuss further the issues I have raised in this letter.
Potentially unconstitutional, the Second Circuit held in Okwedy v. Molinari (2003):
[A] jury could find that Molinari’s letter contained an implicit threat of retaliation if PNE failed to accede to Molinari’s requests. In his letter, Molinari invoked his official authority as “Borough President of Staten Island” and pointed out that he was aware that “P.N.E. Media owns a number of billboards on Staten Island and derives substantial economic benefits from them.” He then “call[ed] on” PNE to contact Daniel L. Master, whom he identified as his “legal counsel and Chair of my Anti-Bias Task Force.”
Based on this letter, PNE could reasonably have believed that Molinari intended to use his official power to retaliate against it if it did not respond positively to his entreaties. Even though Molinari lacked direct regulatory control over billboards, PNE could reasonably have feared that Molinari would use whatever authority he does have, as Borough President, to interfere with the “substantial economic benefits” PNE derived from its billboards in Staten Island.13
Third, the Sheriff of Cook County in Illinois sent letters to Mastercard and Visa saying, “As the Sheriff of Cook County, a father and a caring citizen, I write to request that your institution immediately cease and desist from allowing your credit cards to be used to place ads on websites like Backpage.com [which hosted ads for sex-related services].” Potentially unconstitutional, the Seventh Circuit held in Backpage.com, LLC v. Dart (2015). The court went through the Sheriff’s letter in detail and concluded:
And here’s the kicker: “Within the next week, please provide me with contact information for an individual within your organization that I can work with [harass, pester] on this issue.” The “I” is Sheriff Dart, not private citizen Dart—the letter was signed by “Thomas Dart, Cook County Sheriff.”
And the letter was not merely an expression of Sheriff Dart’s opinion. It was designed to compel the credit card companies to act by inserting Dart into the discussion; he’ll be chatting them up.
Further insight into the purpose and likely effect of such a letter is provided by a strategy memo written by a member of the sheriff’s staff in advance of the letter. The memo suggested approaching the credit card companies (whether by phone, mail, email, or a visit in person) with threats in the form of “reminders” of “their own potential liability for allowing suspected illegal transactions to continue to take place” and their potential susceptibility to “money laundering prosecutions . . . and/or hefty fines.” Allusion to that “susceptibility” was the culminating and most ominous threat in the letter.14
In our fourth and most prominent instance, the Biden administration attempted to persuade social media platforms to block or remove posts on various topics, including “the COVID-19 lab-leak theory, pandemic lockdowns, vaccine side-effects, election fraud, and the Hunter Biden laptop story.” The Fifth Circuit held in Missouri v. Biden (2023) that some of the government’s actions were likely unconstitutionally coercive:
On multiple occasions, the officials coerced the platforms into direct action via urgent, uncompromising demands to moderate content. . . . And, more importantly, the officials threatened—both expressly and implicitly—to retaliate against inaction. Officials threw out the prospect of legal reforms and enforcement actions while subtly insinuating it would be in the platforms’ best interests to comply. As one official put it, “removing bad information” is “one of the easy, low-bar things you guys [can] do to make people like me”—that is, White House officials—“think you’re taking action.” . . . When the officials’ demands were not met, the platforms received promises of legal regime changes, enforcement actions, and other unspoken threats. That was likely coercive. . . .
[M]any of the officials’ asks were “phrased virtually as orders,” like requests to remove content “ASAP” or “immediately.” The threatening “tone” of the officials’ commands, as well as of their “overall interaction” with the platforms, is made all the more evident when we consider the persistent nature of their messages. . . . [T]here is [also] plenty of evidence—both direct and circumstantial, considering the platforms’ contemporaneous actions—that the platforms were influenced by the officials’ demands. . . .
[And] the speaker [had] “authority over the recipient.” . . . [The White House] enforces the laws of our country, and—as the head of the executive branch—directs an army of federal agencies that create, modify, and enforce federal regulations. . . . At the very least, as agents of the executive branch, the officials’ powers track somewhere closer to those of the commission in Bantam Books—they were legislatively given the power to “investigate violations and recommend prosecutions.”
[T]he officials made express threats and, at the very least, leaned into the inherent authority of the President’s office. . . . But, beyond express threats, there was always an “unspoken ‘or else.’” . . . [W]hen the platforms faltered, the officials warned them that they were “[i]nternally . . . considering our options on what to do,” their “concern[s] [were] shared at the highest (and I mean highest) levels of the [White House],” and the “President has long been concerned about the power of large social media platforms.”15
The Supreme Court reversed the Fifth Circuit’s decision on procedural grounds, so that decision is no longer binding precedent.16 The Court’s opinion also cast doubt on the factual findings that the Fifth Circuit relied on.17 Nonetheless, the Fifth Circuit’s analysis may remain persuasive to some judges in future cases.
Of course the coercion/persuasion line is often hazy. One concern about government persuasion of intermediaries is that when the government asks, people who are subject to regulation by the government may hear this as demanding. As it happens, this concern has arisen in at least one other First Amendment context, and the reasoning in that context might be applicable here as well.
That context is labor law. Since the 1940s—early in the Court’s modern First Amendment jurisprudence—the Court has recognized that “employers’ attempts to persuade to action with respect to joining or not joining unions are within the First Amendment’s guarantee” but not when “to this persuasion other things are added which bring about coercion, or give it that character.”18 In NLRB v. Gissel Packing Co. (1969), the Court made clear that the employer’s power over employees should be considered in deciding whether the speech is likely to coerce:
Any assessment of the precise scope of employer expression, of course, must be made in the context of its labor relations setting. . . . [A]ny balancing of [the employer’s and employee’s] rights must take into account the economic dependence of the employees on their employers, and the necessary tendency of the former, because of that relationship, to pick up intended implications of the latter that might be more readily dismissed by a more disinterested ear.19
Similar logic, I think, may apply when high-level executive officials, or those who speak for them, address intermediaries who are regulated by those officials or the officials’ appointees:
[A]ny balancing of [government speakers’ and intermediaries’] rights must take into account the economic dependence of the [intermediaries] on their [regulators], and the necessary tendency of the former, because of that relationship, to pick up intended implications of the latter that might be more readily dismissed by a more disinterested ear.
This analogy would still leave government officials able to make requests in certain ways, just as employers remain able to speak in certain ways to employees about the possible consequences of unionization. But the officials would have to be more careful to make clear that the request carries no threat of retaliation.
There might, however, be arguments that even genuine government persuasion—when there is no coercive threat—aimed at getting social media platforms to remove speech might violate the First Amendment. I’m not sure whether those arguments are ultimately right or wrong, but let me offer a sketch of them.
To begin, let’s consider the Fourth Amendment. Say you rummage through a roommate’s papers, find evidence that he’s committing a crime, and send it to the police. Because you’re a private actor, you haven’t violated the Fourth Amendment. (Whether you committed some tort or crime is a separate question.)20 Because they didn’t perform the search, the police haven’t violated the Fourth Amendment either, and the evidence from this “private search” can be used against the roommate.
But if the police ask you to rummage through the roommate’s papers, that rummaging may constitute a search governed by the Fourth Amendment. “[I]f a state officer requests a private person to search a particular place or thing, and if that private person acts because of and within the scope of the state officer’s request,” then the search would be subject to the constitutional constraints applicable to government searches.21 “Police officers may not avoid the requirements of the Fourth Amendment by inducing, coercing, promoting, or encouraging private parties to perform searches they would not otherwise perform.”22
Indeed, in Skinner v. Railway Labor Executives’ Association (1989), the Supreme Court held that drug tests of railway employees that were authorized but not required by federal regulations were subject to Fourth Amendment scrutiny:
The Government has removed all legal barriers to the testing authorized by Subpart D, and indeed has made plain not only its strong preference for testing, but also its desire to share the fruits of such intrusions. In addition, it has mandated that the railroads not bargain away the authority to perform tests granted by Subpart D. These are clear indices of the Government’s encouragement, endorsement, and participation, and suffice to implicate the Fourth Amendment.23
Considering the extensive regulation of railroads by the government, the railway companies might have felt special pressure to view the government’s “encouragement” and “endorsement” as a command. Yet the Court did not rely on the theory that the government had indeed coerced the railroads to perform the tests. It appeared to be enough that it “encourage[d], endorse[d], and participat[ed]” in the tests. The same may apply to social media platforms, especially (but perhaps not only) in a political environment where there is talk of possible regulation, such as through antitrust law or by modifying Section 230 immunity.24
Likewise, “In the Fifth Amendment context, courts have held that the government might violate a defendant’s rights by coercing or encouraging a private party to extract a confession from a criminal defendant.”25 More broadly, the Supreme Court held in Blum v. Yaretsky (1982), a Due Process Clause case, that “a State normally can be held responsible for a private decision only when it has exercised coercive power or has provided such significant encouragement, either overt or covert, that the choice must in law be deemed to be that of the State.”26 And in Norwood v. Harrison (1973), an Equal Protection Clause case, the Court viewed it as “axiomatic that a state may not induce, encourage or promote private persons to accomplish what it is constitutionally forbidden to accomplish.”27
The inducement, encouragement, and promotion in Norwood involved the provision of tangible benefits (textbooks lent by the state to both public and racially segregated private schools) and not just verbal encouragement. By itself, the line in Norwood thus may not carry much weight. However, the Fourth Amendment cases in which government-encouraged or government-requested private searches became subject to the Fourth Amendment did involve just verbal encouragement.
Though these precedents provide some room for restricting government attempts to persuade platforms to remove speech, and not just attempts at coercion, do such restrictions make sense? After all, government officials have a strong interest in conveying their views, including their views about what speech is harmful and should not be published. It may be that they lack a First Amendment right to do so in their official capacities.28 But there may still be real value to public discourse, and to their listeners, in their being able to do so.
For instance, national security officials might sometimes tell a news outlet, “Look, we can’t force you to do anything, but if you run this story it will lead to deaths of intelligence sources/damage to national security. Could you not run the story, or fuzz over some details, or delay it?” The news outlet might find that to be valuable information. Reporters and editors might want to avoid causing deaths or harming national security, especially if the bulk of the story can still be reported with a bit of delay or slight modification.
Nor is this just some sort of national security exception to a broad presumption that requests not to speak are unconstitutional. Law enforcement officials might reasonably and permissibly tell a newspaper or broadcaster, “If you run this story right now, you’ll tip off the criminals we’re investigating/jeopardize witnesses. Don’t you want us to fight crime effectively?” The newspaper might say yes or no, assuming there’s no context to make the statement coercive. I’m skeptical that this request would violate the First Amendment.
Or say that a newspaper is about to run an op-ed that alleges governmental misconduct. A government official learns of this—perhaps the editors call him to get his side of the story—and says, “That’s nonsense, and here’s the evidence to prove that.” Or he says, “The allegations are so slanted as to be deceptive or unfair; here’s the context that shows it.” And then adds, “Please don’t run such an unfair story; it would be bad for us if you did, but it would also be bad for your reputation, when the truth comes out, and it would be bad for your readers, who would be misled.”
That is a call for an intermediary (the newspaper) to block the publication of a third-party item (the op-ed). However, it is unlikely to be unconstitutional. Indeed, the newspaper may be quite pleased to learn the full story and thereby avoid publishing an op-ed that would make the newspaper look bad.
Perhaps, though, one difference might be between occasional one-off conversations and systematic programs. To be sure, when it comes to coercive threats aimed at suppressing speech, both the ad hoc and systematic demands are unconstitutional.29 Likewise, the cases involving government encouragement of searches by private parties find even ad hoc demands unconstitutional.30
But if courts do conclude that ad hoc requests to remove or block speech are constitutional, perhaps some line should still be drawn between those requests and systematic encouragement of such removing or blocking. This appears to be what the Fifth Circuit concluded in Missouri v. Biden, when it found that the government’s speech was impermissible “significant encouragement” of speech restriction by platforms, even apart from the coercion argument:
The officials had consistent and consequential interaction with the platforms and constantly monitored their moderation activities. In doing so, they repeatedly communicated their concerns, thoughts, and desires to the platforms. The platforms responded with cooperation—they invited the officials to meetings, roundups, and policy discussions. And, more importantly, they complied with the officials’ requests, including making changes to their policies. . . .
When the platforms’ policies were not performing to the officials’ liking, they pressed for more, persistently asking what “interventions” were being taken, “how much content [was] being demoted,” and why certain posts were not being removed. Eventually, the officials pressed for outright change to the platforms’ moderation policies. . . . Beyond that, they relentlessly asked the platforms to remove content, even giving reasons as to why such content should be taken down. They also followed up to ensure compliance and, when met with a response, asked how the internal decision was made. . . .
Consequently, it is apparent that the officials exercised meaningful control—via changes to the platforms’ independent processes—over the platforms’ moderation decisions. By pushing changes to the platforms’ policies through their expansive relationship with and informal oversight over the platforms, the officials imparted a lasting influence on the platforms’ moderation decisions without the need for any further input. In doing so, the officials ensured that any moderation decisions were not made in accordance with independent judgments guided by independent standards. Instead, they were encouraged by the officials’ imposed standards.
In sum, we find that the White House officials, in conjunction with the Surgeon General’s office, coerced and significantly encouraged the platforms to moderate content. As a result, the platforms’ actions “must in law be deemed to be that of the State.”31
Indeed, when it came to requests for removal made by the Centers for Disease Control and Prevention, the Fifth Circuit concluded that the requests were not coercive, but still constituted unconstitutional significant encouragement:
[T]he CDC was entangled in the platforms’ decision-making processes. The CDC’s relationship with the platforms began by defining—in “Be On the Lookout” meetings—what was (and was not) “misinformation” for the platforms. Specifically, CDC officials issued “advisories” to the platforms warning them about misinformation “hot topics” to be wary of. From there, CDC officials instructed the platforms to label disfavored posts with “contextual information,” and asked for “amplification” of approved content. That led to CDC officials becoming intimately involved in the various platforms’ day-to-day moderation decisions. For example, they communicated about how a platform’s “moderation team” reached a certain decision, how it was “approach[ing] adding labels” to particular content, and how it was deploying manpower. Consequently, the CDC garnered an extensive relationship with the platforms.
From that relationship, the CDC, through authoritative guidance, directed changes to the platforms’ moderation policies. . . . [The platforms] adopted rule changes meant to implement the CDC’s guidance. . . . Thus, the resulting content moderation, “while not compelled by the state, was so significantly encouraged, both overtly and covertly” by CDC officials that those decisions “must in law be deemed to be that of the state.”32
As noted above, the Supreme Court reversed this Fifth Circuit decision on procedural grounds and cast some doubt on the factual findings on which the Fifth Circuit relied.33 But the Fifth Circuit’s legal analysis as to substantial encouragement and systematic entanglement may remain persuasive to lower courts.
Of course, distinguishing “consistent and consequential interaction” from mere occasional interaction—such as the examples of constitutionally permissible requests given above—can be difficult. Still, constitutional law does sometimes draw such distinctions between occasional action and systemic action. One analogy, though distant, might be how the law sometimes treats administrative searches.
Courts have upheld various kinds of searches—even ones that lack a warrant, probable cause, or both—on the grounds that they are targeted at specific public safety concerns rather than at broad law enforcement. Airport searches of luggage, aimed at detecting weapons, are one example, as the Ninth Circuit discussed in detail in United States v. $124,570 U.S. Currency (1989).34
Now say that Transportation Security Administration agents, U.S. government employees following their normal duty to search for weapons, spot a suspicious amount of cash or drugs. They then alert the police who use this information as part of the probable cause needed to justify a search. That is constitutional.35 TSA agents are free to “report information pertaining to criminal activity, as would any citizen.”36
So far, so good. But say that the Drug Enforcement Administration comes up with a systematic program to encourage TSA agents to search not just for weapons, the rationale that led airport searches to be upheld in the first place, but also for drugs or cash. The Ninth Circuit held that this would be going too far:
We see the matter as materially different where the communication [about the drugs or money that the TSA agent found] is undertaken pursuant to an established relationship, fostered by official policy, even more so where the communication is nurtured by payment of monetary rewards.37
Even if ad hoc reporting by TSA agents to the police of things other than weapons is permissible under the Fourth Amendment, a system set up to encourage such reporting is not. “The line we draw is a fine one but, we believe, one that has constitutional significance.”38
Or consider sobriety checkpoints. The Court has upheld them as permissible administrative seizures because they are aimed at protecting safety on the very roads that are being temporarily blocked.39 Yet the Court has held that the government may not set up drug trafficking checkpoints aimed at finding drug dealers.40 The difference in these cases, the Court held, stems from the “difference in the Fourth Amendment significance of highway safety interests and the general interest in crime control.”41
Now, if officers conducting sobriety checkpoints happen to see evidence of crime in plain sight—blood on the seat, an illegally carried gun, or, for that matter, drugs—they are free to keep detaining the driver and search further, based on this newly discovered probable cause.42 But say that the checkpoint is set up precisely for this purpose, as a systematic way of searching for drugs or for other contraband. That would trigger additional Fourth Amendment scrutiny: ad hoc observation of evidence of crime, in the course of a valid administrative seizure (valid because the seizure is part of a drunk drivingcheckpoint, rather than a drug checkpoint or a general law enforcement checkpoint), may become unconstitutional if it happens in the course of a systematic program of search for evidence of crime.43
I should stress again that these analogies are imperfect. Among other differences, they involve the Fourth Amendment and not the First, and concern attempts to systematically encourage certain action by government employees and not by private parties. But my point here is that they offer some support for the view that even if some actions are not subject to constitutional scrutiny when done on a one-off basis, they may become unconstitutional when done systematically. In the Fourth Amendment context, systematizing permissible ad hoc searches into “an established relationship, fostered by official policy” increases the threat of undue government intrusion on privacy, enough to change the Fourth Amendment analysis. Perhaps systematizing permissible ad hoc requests not to publish something into a similar official established relationship may likewise increase the threat of undue government interference with public debate to the point that First Amendment scrutiny would be required.
Perhaps. Maybe judicial line drawing here is so difficult that any result ought to be implemented by Congress rather than by courts. I’m not sure what the right ultimate result ought to be. Still, the analogies may be useful for thinking through the question.
These questions also expose one interesting feature of state laws that restrict the ability of social media platforms to act as intermediaries controlling user speech.44 Texas and Florida have passed two such statutes, and other states may do the same; the Supreme Court held that they unconstitutionally interfered with the platforms’ ability to choose what goes in their “curated feeds” (such as Facebook’s news feed), but didn’t resolve whether the statutes might limit platforms’ ability to deplatform users or remove individual user posts.45
The laws are generally billed as attempts to protect users from undue viewpoint discrimination by platforms, and they are often criticized as unduly restricting the rights of these platforms. Yet it’s worth noting that, by stripping platforms of the power to yield to government encouragement or subtle coercion, the laws also functionally strip federal and state executive branches of power from engaging in such encouragement or coercion.
If the U.S. government had federal statutory authority to order platforms to remove certain posts, then state laws would be preempted by that hypothetical federal statute. But if the federal government claims that it is not ordering platforms to do something but is merely asking the platforms to exercise their own powers instead, then state law can indeed stymie such federal requests by forbidding platforms from exercising their powers in that way.46
Of course, one can argue that the laws throw the baby out with the bathwater. Even if one thinks that government encouragement of private platforms’ speech restrictions is generally improper, one can conclude that these platforms’ own decisions to restrict speech are just fine—indeed, that such voluntary private decisions are constitutionally protected exercises of editorial discretion, and possibly great contributions to the public good. Yet, in weighing the costs and benefits of these laws, one possible benefit is that they end up limiting government power to control public debate in the process of limiting platform power.47
The internet has democratized speech, restricted the power of one set of intermediaries (traditional media), and empowered a new set (social media platforms). In the process, it has made the latter tempting targets for government persuasion and pressure. The future of government efforts to restrict online speech will likely continue to include a great deal of both persuasion and pressure. What the law should say about such government action is an increasingly important topic of debate.
Endnotes
- 1A. J. Liebling, The Press (New York: Ballantine Books, 1975), 32.
- 2For a mostly optimistic discussion of cheap speech that also notes “a possible dark side,” see Eugene Volokh, “Cheap Speech and What It Will Do,” The Yale Law Journal 104 (7) (1995): 1805–1850. At the time, it, like this volume, was a work of futurism. For a review of the article as futurism, see Matt Novak, “” Paleofuture, June 26, 2014. Novak states, for example, “In 1995 Eugene Volokh wrote . . . an incredibly prescient meditation on the future of media and technology. But it has just enough weird anachronisms to remind us that nobody can predict the future with absolute certainty. Think of it as the uncanny valley of old futurism—so incredibly close to the future that actually arrived, but just inaccurate enough that it gives you a weird feeling in the pit of your stomach.”
- 3See Jack M. Balkin, “How to Regulate (and Not Regulate) Social Media,” Journal of Free Speech Law 1 (1) (2021): 71–96; Ashutosh Bhagwat, “Do Platforms Have Editorial Rights?” Journal of Free Speech Law 1 (1) (2021): 97–138; Eric Goldman and Jess Miers, “Online Account Terminations/Content Removals and the Benefits of Internet Services Enforcing Their House Rules,” Journal of Free Speech Law 1 (1) (2021): 191–226; Mark A. Lemley, “The Contradictions of Platform Regulation,” Journal of Free Speech Law 1 (1) (2021): 303–336; Eugene Volokh, “Treating Social Media Platforms Like Common Carriers?” Journal of Free Speech Law 1 (1) (2021): 377–462; Adam Candeub and Eugene Volokh, “Interpreting 47 U.S.C. § 230(c)(2),” Journal of Free Speech Law 1 (1) (2021): 175–190; and Jack Goldsmith and Eugene Volokh, “State Regulation of Online Behavior: The Dormant Commerce Clause and Geolocation,” Texas Law Review 101 (5) (2023): 1083–1125. For an earlier and broader perspective on statutes that aim to increase people’s free speech rights (though perhaps at the expense of other private entities), see Genevieve Lakier, “The Non–First Amendment Law of Freedom of Speech,” Harvard Law Review 134 (7) (2021): 2299–2381.
- 4See Eugene Volokh, “The Reverse Spider-Man Principle: With Great Responsibility Comes Great Power,” Journal of Free Speech Law 3 (1) (2023): 197–216.
- 5The leading work on the subject is Derek E. Bambauer, “Against Jawboning,” Minnesota Law Review 100 (1) (2015): 51–126. For a notable new article on the subject, see Philip Hamburger, “Courting Censorship,” Journal of Free Speech Law 4 (1) (2024): 195–298.
- 6603 U.S. ___, 144 S. Ct. 1972 (2024).
- 7707 F.2d 33, 34, 37, 38–39 (2d Cir. 1983).
- 8939 F.2d 1011, 1013, 1015–1016 (D.C. Cir. 1991).
- 9196 F.3d 56, 68, 70 (2d Cir. 1999).
- 10372 U.S. 58 (1963).
- 11602 U.S. 175, 197 (2024). Note that I was one of the NRA’s lawyers in this case.
- 12930 F.2d 204, 209–210 (2d Cir. 1991).
- 13333 F.3d 339, 341, 342, 344 (2d Cir. 2003).
- 14807 F.3d 229, 231–32 (7th Cir. 2015). The bracketed words, “harass, pester,” were added by the court, presumably as an indication of how the court interpreted “work with.” See Complaint Exh. B at 7, Backpage.com, LLC v. Dart, No. 1:15-cv-06340 (N.D. Ill. July 21, 2015).
- 1583 F.4th 350, 382 (5th Cir. 2023).
- 16603 U.S. ___, 144 S. Ct. 1972 (2024).
- 17Ibid., n.4.
- 18Thomas v. Collins, 323 U.S. 516, 537 (1945) (treating the matter as having been settled by NLRB v. Virginia Electric & Power Co., 314 U.S. 469 [1941]); Virginia Electric & Power, 314 U.S. at 477 (“The employer in this case is as free now as ever to take any side it may choose on this controversial issue. But, certainly, conduct, though evidenced in part by speech, may amount, in connection with other circumstances, to coercion within the meaning of the Act”).
- 19395 U.S. 575, 617 (1969).
- 20See United States v. Phillips, 32 F.4th 865, 867 (9th Cir. 2022); and Burdeau v. McDowell, 256 U.S. 465, 475–476 (1921).
- 21State v. Tucker, 330 Or. 85, 90 (2000) (applying the Oregon Constitution’s Fourth Amendment analog; police request to tow truck driver to search items in car being towed), followed by State v. Lien, 364 Or. 750, 778 (2019) (police request to trash company to pick up a person’s trash in a particular way that would facilitate its being searched). See also United States v. Gregory, 497 F. Supp. 3d 243 (E.D. Ky. 2020) (similar fact pattern to Lien).
- 22George v. Edholm, 752 F.3d 1206, 1215 (9th Cir. 2014) (police request to doctor to do a rectal search). See also United States v. Ziegler, 474 F.3d 1184, 1190 (9th Cir. 2007) (police request to employer to search employee’s work computer); and United States v. Rosenow, 50 F.4th 715, 733 (9th Cir. 2022) (recognizing that, even when a private party’s search would normally be entirely legal, the government’s “encouragement” of such a search may constitute “state action”).
- 23Skinner v. Railway Labor Executives Association, 489 U.S. 602, 615–616 (1989).
- 24See Murthy v. Missouri, 603 U.S. ___, 144 S. Ct. 1972 (2024) (Samuel Alito, dissenting) (reasoning that “internet platforms, although rich and powerful, are at the same time far more vulnerable to Government pressure than other news sources” because “[t]hey are critically dependent on the protection provided by § 230 of the Communications Decency Act of 1996,” which Congress might threaten to withdraw; “[t]hey are vulnerable to antitrust actions”; and, “because their substantial overseas operations may be subjected to tough regulation in the European Union and other foreign jurisdictions, they rely on the Federal Government’s diplomatic efforts to protect their interests”).
- 25United States v. Folad, 877 F.3d 250, 253 (6th Cir. 2017) (emphasis added). See also United States v. Garlock, 19 F.3d 441, 443–444 (8th Cir. 1994).
- 26457 U.S. 991, 1004 (1982). See also Fidelity Financial Corp. v. Federal Home Loan Bank of San Francisco, 792 F.2d 1432, 1435 (9th Cir. 1986) (emphasis added).
- 27413 U.S. 455, 465 (1973) (emphasis added).
- 28Compare Garcetti v. Ceballos, 547 U.S. 410, 421 (2006) (concluding that government officials generally don’t have First Amendment rights when exercising their official duties); and David Fagundes, “State Actors as First Amendment Speakers,” Northwestern University Law Review 100 (4) (2006): 1637–1688 (discussing uncertainty about when state officials may have First Amendment rights vis-à-vis the federal government).
- 29See Fagundes, “State Actors as First Amendment Speakers,” Part II.B.
- 30See ibid., Part IV.A.
- 3183 F.4th 350, 387 (5th Cir. 2023).
- 32Ibid., 390.
- 33See 603 U.S. ___, 144 S. Ct. 1972 (2024).
- 34873 F.2d 1240, 1244–1245 (9th Cir. 1989).
- 35See ibid., 1247 n.7 (approvingly describing United States v. Canada, 527 F.2d 1374, 1376, 1378–1379 [9th Cir. 1975]).
- 36Ibid.
- 37Ibid.
- 38Ibid.
- 39Michigan v. Sitz, 496 U.S. 444 (1990).
- 40City of Indianapolis v. Edmond, 531 U.S. 32 (2000).
- 41Ibid., 40.
- 42See Texas v. Brown, 460 U.S. 730, 744 (1983) (plurality opinion); ibid., 746 (Lewis F. Powell Jr., concurring in the judgment); and People v. Edwards, 101 A.D.3d 1643, 1644 (2012).
- 43I borrow this from United States v. Soyland, 3 F.3d 1312, 1317 (9th Cir. 1993).
- 44Volokh, “Cheap Speech and What It Will Do.”
- 45Moody v. Netchoice, LLC, 603 U.S. ___, 144 S. Ct. 2383 (2024).
- 46By way of analogy, say that a state law bans private employers from firing employees for their political speech, as many state laws do. SeeEugene Volokh, “Private Employees’ Speech and Political Activity: Statutory Protection Against Employer Retaliation,” Texas Review of Law & Politics 16 (295) (2012): 295–336. Then if a federal official urges a company in that state to fire someone for that person’s political views, the company will be forbidden from doing so, and any pressure the federal official seeks to exert will thus be ineffective.
- 47Compare Ian Samuel, “The New Writs of Assistance,” Fordham Law Review 86 (6) (2018): 2873–2924 (similarly arguing that network service providers should be limited in their ability to gather information about users, because of the concern that the government will co-opt those providers into part of the government’s own surveillance system).