Notice & Comment

The First Amendment Does Not Prohibit The Government From Addressing Big Tech Censorship, by Brendan Carr and Nathan Simington

I. Introduction. 

The Supreme Court has described social media as “the modern public square.”[1]  And today, a small group of some of the world’s most powerful corporations exercise complete control over that public square.  Big Tech decides who gets to speak, what they can say, and who gets to listen to that speech.  This level of concentrated power over a central means of communication has an outsized impact on public discourse and political debate in this country.  Indeed, social media censorship is imposing real-world harms on Americans.  There is now an abundance of evidence that these companies’ decisions to discriminate against other peoples’ speech have prevented Americans from accessing and discussing critically important information, including in times of crisis.[2]

In response, several states have taken actions that enable a diversity of viewpoints to compete in the marketplace of ideas.  Texas, for example, enacted HB20, a law that prohibits dominant social media platforms from engaging in viewpoint-based discrimination and requires them to disclose truthful information about their services.[3] Notably, HB20 imposes no constraint whatsoever on Silicon Valley’s own speech.  These companies remain free to decide for themselves whether they want to speak on any topic (and, if so, what they will say) or remain silent altogether.  What HB20 does is permissibly regulate the ability of Big Tech to control whether other people can speak and participate in the digital public square.  And HB20 does so in a way that will promote more speech, not less.  The Supreme Court will consider the constitutionality of HB20 in NetChoice, LLC v. Paxton later this year.  NetChoice, a consortium of the largest online platforms, does not seriously dispute the scope and scale of its viewpoint-based censorship.  Instead, these companies propound a novel and sweeping theory of the First Amendment that conflicts with Supreme Court precedent.

NetChoice broadly asserts that the First Amendment puts its decisions to censor or otherwise discriminate against other peoples’ speech beyond the reach of the government.  In doing so, it embraces Supreme Court cases like Tornillo, which involved a state law that compelled newspapers to run partisan editorials, and similar precedents like Pacific Gas & Electric and Hurley.  But as we discuss below, NetChoice’s reliance on that line of authorities is misplaced both as a matter of law and fact.  The Supreme Court has never held that the First Amendment grants dominant companies like social media giants a freewheeling right to censor others’ speech.

To the contrary, the Court has recognized that the government can apply antidiscrimination requirements—including ones that condition the right of a private entity to exclude other people and their speech—to corporations consistent with the First Amendment in appropriate cases.  The Court has upheld laws like that in many instances, including Turner (requiring that cable systems carry broadcast television channels), Rumsfeld v. FAIR (requiring law schools to allow military recruiters equal access to their campuses as a funding condition), Pruneyard (requiring a private owned shopping mall to allow political petitioning inside its property), and Zauderer (requiring disclosures in lawyer advertising). 

Social media laws like HB20 are easily distinguished from those struck down on First Amendment grounds in Tornillo and its progeny.  Indeed, HB20 touches none of the First Amendment third rails that were at play in those cases.  Instead, HB20 fits comfortably within the body of laws upheld by the Supreme Court in cases like TurnerFAIRPruneyard, and Zauderer.   

Properly read, the relevant Supreme Court cases largely rely on the presence or absence of certain factors.  Those include market considerations (like market power as well as the centrality of the regulated medium to public discourse); the degree to which the regulated entity makes individualized decisions about what speech to carry instead of operating mainly as a conduit for other peoples’ speech; whether the challenged law compels a response; and whether the law’s requirements are triggered by the content of speech or risk attributing speech to the complaining party.  When viewed through the lens of these handful of key factors, it is clear that the government can, in the appropriate case, apply antidiscrimination rules to social media platforms.  Texas’s HB20 is one of those cases. 

We base our analysis on our experience as Commissioners of the Federal Communications Commission (FCC), an independent agency charged by Congress with promoting a vibrant and diverse media ecosystem consistent with the First Amendment.  The FCC administers antidiscrimination protections in a variety of contexts, from cable television to broadcast transmission to common carrier services like traditional telephone offerings.  And the FCC oversees the Communications Act—which includes Section 230, the liability shield that Big Tech regularly invokes to escape any regulation.  As FCC Commissioners, we are deeply concerned that Big Tech’s expansive position threatens free speech and is inconsistent with the nation’s communications regulatory framework. 

At the same time, upholding HB20 would not pave the way for indiscriminate regulation of social media or any other communications medium.  Take the FCC’s 2015 decision to apply common carrier regulations to Internet Service Providers (ISPs).  We are very familiar with this regulatory regime because we both voted to dissent from the FCC’s 2023 decision to head back down the path of reinstating those regulations.[4]  There are several reasons why the FCC’s 2015 net neutrality rules failed First Amendment scrutiny that are not present in this social media context.  Indeed, then-Judge Kavanaugh presciently identified some of those distinctions in his 2017 decision to dissent from the D.C. Circuit’s denial of rehearing en banc in the net neutrality case.  

First, as then-Judge Kavanaugh noted, the FCC’s 2015 decision included no finding of market power.  There is market power here.  Second, the FCC could not have made the requisite market power finding even if it had tried.  Indeed, there are key differences between the ISP and social media markets that matter for the First Amendment analysis.  Third, the FCC failed to offer up any valid government interest for those net neutrality rules.  Fourth, the FCC’s 2015 net neutrality rules are distinct from HB20 because they operated as an intrusive and overbroad set of regulations.  Fifth, there is a state action overlay when it comes to Big Tech censorship (see Missouri v. Biden) that is not and was not present in the ISP context.  Sixth, NetChoice’s unique invocation of Section 230—which makes a legal distinction between platforms and the third-party speech they host—further undermines its First Amendment argument. [5]

At bottom, ruling in favor of Big Tech in NetChoice would effectively require the Supreme Court to disregard decades of First Amendment precedent and throw asunder legions of well-established antidiscrimination laws.  In contrast, ruling in favor of upholding Texas’s social media law would not only be consistent with the Supreme Court’s First Amendment cases but would not compel any court to uphold overbearing laws like the FCC’s 2015 net neutrality regulation.

II. NetChoice’s Sweeping Legal Theory of the First Amendment Is Legally Unsupported.

NetChoice claims that it has an unbridled right to censor or otherwise discriminate against other peoples’ speech.  That claim is false, as we discuss below.  In Part II.A, we note that this claim runs headlong into Supreme Court precedent, which establishes that a private party’s right to exclude other people or their speech can be regulated by the government in appropriate cases.  In Part II.B, we review some of the Supreme Court’s most relevant First Amendment precedents, and we identify a handful of factors that appear pivotal to the Court’s analysis.  Application of those factors in the context of social media companies point in favor of upholding reasonable social media laws like Texas’s HB20.

A. The First Amendment Does Not Give Dominant Providers the Unqualified Right to Censor Speech. 

NetChoice has broadly challenged antidiscrimination laws on the ground that the First Amendment prohibits the government from ever regulating the decisions that a private entity makes to exclude or refuse to host the speech of other people.  That sweeping claim is incorrect.  

The Supreme Court’s decisions confirm that a private entity’s right to exclude the speech of others is not unlimited.  Indeed, the Court has permitted certain safeguards where the regulated entity exercises market power and controls a central means of communication, where the speech or views of the regulated entity will not be confused with those of the private interests whose speech is enabled, or where the exception is limited and tied closely to a state interest to protect consumers. 

As an initial matter, the First Amendment generally presents no barrier to antidiscrimination rules applied to common carriers like telephone companies, railroads, and postal services.[6]  Even outside the context of common carriers, the First Amendment does not operate as a complete bar to all regulations.  In the context of communications networks, for instance, the Supreme Court in Turner upheld federal must-carry rules for cable television operators.[7]  Those rules required cable television operators to devote some of their channels for the carriage of local broadcast programming.  The Court rejected the argument that those must-carry rules violated the First Amendment by compelling cable operators to transmit speech not of their choosing.  Rather, the rules were justified, the Court determined, by special characteristics of the cable medium, and they were not structured in a way that would undermine First Amendment interests.[8]

Outside communications networks, courts have also permitted limited intrusions on the right to exclude speech from private property or platforms.  In Pruneyard, the Court found that allowing individuals to exercise their own free speech and petition rights on the grounds of a privately owned shopping center did not violate the First Amendment rights of the property owner.[9]  In rejecting the owner’s purported right to be free from using his property as a forum for the speech of others, the Court noted that the shopping center was a business open to the public, whose speech would not be confused with that of the owner.[10]  In Zauderer, the Court upheld a state disclosure requirement for fee arrangements in commercial advertising by attorneys.  These disclosures, the Court reasoned, were “reasonably related to the State’s interest in preventing deception of consumers.”[11]  

To be sure, the Court has struck down certain laws that compel the speech of private actors.  In Tornillo, the Court invalidated a state law requiring any newspaper that assailed a political candidate’s character to print, upon request by the candidate and without cost, the candidate’s reply in equal space and prominence.[12]  Likewise, in Pacific Gas & Electric, the Court invalidated a law requiring a private utility to include with its monthly bills a newsletter published by a consumer group critical of the utility’s ratemaking practices.[13]  Finally, in Hurley, the Court invalidated application of a state antidiscrimination law that would have required a parade to include participation by an advocacy organization.[14]  

Given the Court’s myriad cases to the contrary, however, TornilloPacific Gas & Electric, and Hurley do not establish an absolute right against carrying speech on nondiscriminatory terms.  As discussed below, moreover, they do not vindicate NetChoice’s broader legal theory about “compelled speech,” and are distinguishable from antidiscrimination protections like HB20 on the facts.  

B. The First Amendment Analysis Holistically Incorporates Multiple Factors Instead of a Simplistic “Compelled Speech” Dichotomy. 

We have distilled a handful of factors from the Supreme Court’s First Amendment precedents that seem particularly relevant for the Court’s consideration in these types of cases.  The presence or absence of those factors is not necessarily outcome determinative, but they can help inform whether a particular regulation is likely to be upheld or invalidated.  These factors include: 

  • Does the regulated entity have market power?  Is its platform central to public discourse?  
  • Is the regulated entity a passive conduit that holds itself open to all third-party messages?  Or is it in the business of prescreening messages and making individualized decisions about which speech to carry?  
  • Is the message reasonably attributable to the regulated entity?  Will the regulated entity feel pressured to reply or distance itself from the message? 
  • Does the law at issue require carriage of all messages from third parties, or just a particular kind of message?  

When viewed through these key First Amendment factors, regulations like HB20 fall comfortably within the types of the antidiscrimination regulations the Court has upheld.  Considering the issues at play in these varying cases, HB20 looks a lot more like the laws upheld by the Court in TurnerPruneyard, and Zauderer than the unconstitutional requirements at issue in TornilloHurley, and Pacific Gas & Electric.  We discuss each factor below and explain how they weigh in favor of upholding the reasonable regulation of social media companies.

i. Social Media Platforms Exercise Market Power and Are Central to Modern Public Discourse.

Social media platforms compare themselves to traditional newspapers to claim that strict scrutiny applies.  That superficial analogy glosses over the relevant characteristics of the medium at issue.  Importantly, First Amendment doctrine considers the market power of a regulated entity and its centrality to civic engagement—factors that are present in a dominant social media platform but not in a newspaper.   

Take Turner, where cable operators characterized themselves like newspapers to escape the obligation to carry broadcast programming.  The Court rejected that analogy, however, because a single cable operator in the 1990s wielded unilateral access over video programming in a way the newspaper did not—and could not—over print journalism.[15]  The Court found that a cable operator in the 1990s controlled an “essential pathway” for video programming through its physical facilities.[16]  While cable is “not subject to regulation as a common carrier or utility,”[17] it was still found to exercise “bottleneck, or gatekeeper, control” over a significant portion of programming available to the residential consumer.[18] Its control over this essential distribution medium gave it power at the time to silence competing speech instantaneously, overriding the natural check of market forces.[19] A single newspaper, no matter its market share, lacks similar gatekeeper power when it does not control access to all or substantially all competing print media.[20]  These differences, the Court reasoned, gave the government greater leeway to promote the free flow of information by regulating 1990s cable.[21]  In Hurley, the Court distinguished parades from cable operators, which were a “conduit for speech produced by others” with the “monopolistic opportunity to shut out some speakers.”[22]

These characteristics apply to dominant social media platforms today.  As the Supreme Court has recognized, they are “for many … the principal sources for knowing current events, checking ads for employment, speaking and listening in the modern public square, and otherwise exploring the vast realms of human thought and knowledge.”[23]  The economic structure of social media platforms makes this phenomenon unavoidable.  They are walled gardens inherently vulnerable to concentration and have shown remarkable immunity from new forms of competition.  The value of a platform depends on its size—the more users it has, the more vital it becomes.  And once a platform achieves enough scale, it becomes almost irreplaceable to the public at large.  Dominant platforms can represent the outcome of a “winner-take-all” dynamic that is characterized by network effects, switching costs, the self-reinforcing advantages of unique data sets, and increasing returns to scale.  The barriers to competitive entry are significant, and dominant platforms lock in their users from viable alternatives, who in turn often have no realistic substitute to access content outside the walled garden.  

While Facebook, YouTube, and other platforms are dominant, each is dominant in its own way.  They provide distinct services, rather than operate as direct head-to-head competitors, and treating them as complete substitutes does not reflect the realities of consumer behavior or their influence in the marketplace of ideas.  For example, Twitter (now X) broadly speaking never had a dominant market share of social media users or online advertising, but 69% of journalists considered it to be a top platform used for their jobs.[24]

ii. Social Media Platforms Operate in the Main as Conduits for Other Peoples’ Speech.

Another factor emerging from First Amendment precedent is the degree to which the regulated entity makes individualized decisions about the speech it carries—as opposed to operating in the main as a conduit for other peoples’ speech.  The Supreme Court has been far more likely to strike down laws that regulate entities whose business it is to make individualized content decisions.  This factor cuts in favor of upholding HB20’s application to the dominant social media companies.

The 1990s-era cable systems again serve as a useful starting point.  They served dual roles: while they sometimes developed their own programming, in reality they were in the business of passively carrying programming run by unaffiliated channels.  Much of that programming was purchased by those channels from production companies after the cable carriage agreements had been signed, meaning that the cable companies were often contractually bound to carry third-party content that did not even exist when they agreed to carry a channel.  In other words, cable operators did not prescreen the material or exercise individualized control over the creation of programs developed by third parties.[25]  

A newspaper, on the other hand, is not merely a “passive conduit or receptacle.”[26]  Newspapers are in the business of determining what content to run, where it goes on the page, and how long it can be.  The characteristics of the medium essentially makes all speech printed in a newspaper the newspaper’s own speech.  And critically, a newspaper differentiates itself by prescreening, selecting, and curating the content that it disseminates, especially on public figures and social issues.[27]  When laws impinge on that function, as was the case in Tornillo, the Supreme Court has applied strict scrutiny.

Here, dominant social media platforms do not normally control third-party content in the way that newspapers do.  While a social media platform may produce some of its own content or editorialize on the third-party content it delivers, that activity is distinct from their transmission of user-generated content.  The latter—which is at issue here—is the core business of a platform and it brings the regulated conduct much closer to Turner than Tornillo.  Like cable television in the 1990s, user-generated social media content is largely transmitted—nearly instantaneously—on a “continuous and unedited basis to subscribers.”[28]  

iii. User Speech Is Not Reasonably Attributable to Social Media Platforms.

Big Tech appears to argue in NetChoice and similar contexts that user speech is reasonably attributable to the companies that run the platforms.  That apparent endorsement, the argument goes, puts the platforms on stronger First Amendment footing when they push back on laws like HB20.  Section 230 disposes of that argument in the first instance: as a matter of law, platforms are not “the publisher[s] or speaker[s]” of the third-party speech they carry, because they were found to be a “forum for a true diversity of political discourse.”[29]  In contrast, Tornillo emphasized newspapers are often “intensely partisan and narrow in their views” and have “never been entirely free from liability for what it chooses to print.”[30]

Section 230 aside, HB20 differs from laws that pose attribution or endorsement concerns.  Recall the right-of-reply law invalidated in Tornillo.[31]  Because of the unique nature of the newspaper medium, its readers would naturally assume that the newspaper endorsed the reply or otherwise made an affirmative decision to run it.  Similarly, the law invalidated in Pacific Gas & Electric heightened the risk that the utility was perceived as endorsing the critical views it published.[32]  

Not so with social media.  Because of how the medium operates, with posts being attributed to users and transmitted instantly without screening, readers have no reason to assume that the platform chose not to censor a particular post because the platform agreed with the message.  The platforms’ own practices—allowing Ayatollah Khamenei to call for the destruction of Israel and cheer the killing of Jews while censoring American conservatives for “disinformation”—belie their purported concerns about attribution or endorsement.  Like in Turner[33] and FAIR[34] HB20 poses very little risk that an observer would assume that user content—which identifies the actual speaker—is being endorsed by the digital platform.  

iv. Laws that Do Not Discriminate Against Viewpoints Are More Likely to Survive First Amendment Scrutiny.

NetChoice asserts that antidiscrimination regulation of platform moderation warrants strict scrutiny.  But as relevant here, Supreme Court precedent awards the highest level of scrutiny only to regulations that “suppress, disadvantage, or impose differential burdens upon speech because of its content”[35] or those that have compelled speech that is “bearing a particular message.”[36]  In contrast, the Court applies intermediate scrutiny when challenged regulations are unrelated to the content of speech.[37]  

As an example, the broadcast carriage laws at issue in Turner were content neutral because, in the Court’s view, they did not depend on the programming’s content.[38]  Likewise, the law at issue in FAIR did not “dictate the content of the speech at all” by tying federal funding and campus access to military recruiting, and in any case created no risk that observers would associate the military’s speech with the universities.[39]  Tornillo, on the other hand, involved a content trigger: the newspaper was forced to run a reply only after it printed criticism of a political candidate.[40]  Pacific Gas & Electric similarly involved a content-based distinction: it required utilities to run the critical speech of a single consumer group.  While Pacific Gas & Electric did not deal with a newspaper, the Court applied the same framework and analogized the bill insert to a newspaper.[41]  It is therefore unsurprising that Tornillo and Pacific Gas & Electric did not control the outcome in Turner and FAIR.

The same outcome should follow here.  The government can craft antidiscrimination rules that lawfully address viewpoint discrimination.  HB20 does just that.  It does not “distinguish[] favored speech from disfavored speech on the basis of the ideas or views expressed.”[42]  Instead, it operates to protect all viewpoints and aims to ensure there are more ideas or viewpoints in the public discourse. 

III. Rejecting NetChoice’s Sweeping Legal Theory Does Not Open the Door to Indiscriminate Regulation.

There is much to say for judicial modesty, and NetChoice’s expansive theory of “compelled speech” reaches beyond the instant case to threaten a wide array of well-settled antidiscrimination regulations.  At the same time, one should be mindful of the opposite risk—namely, that upholding a regulation might open the door to indiscriminate regulation of communications mediums and networks.  The Supreme Court, fortunately, is not faced with that false choice in the case of HB20.  The Supreme Court could easily uphold the Texas law while preserving legitimate First Amendment claims against government overreach.  

Compare HB20 to the FCC’s 2015 net neutrality rules, which applied utility-style regulations to competitive broadband networks for the first time in American history.[43]  As noted above, the two of us are very familiar with the details of that FCC regime because we recently voted against the FCC majority’s decision to start the process of re-imposing those burdensome rules.  When the FCC last adopted similar regulations in 2015, the D.C. Circuit, en banc, upheld them.  But then-Judge Kavanaugh found those unprecedented rules fatally infirm under the First Amendment.  As discussed below, there is no conflict at all between then-Judge Kavanaugh‘s dissent in that case and a decision by the Supreme Court to uphold HB20 against NetChoice’s First Amendment claims.  Indeed, Judge Kavanaugh’s dissent presciently identified some of the key differences between impermissible rules, like the FCC’s 2015 net neutrality regulations, and lawful enactments, like HB20.  

First, as Judge Kavanaugh noted in his 2017 D.C. Circuit dissent, the FCC in 2015 made no finding of market power to justify its net neutrality rules.  Indeed, the FCC did not even try.  If it had done so successfully, Judge Kavanaugh wrote, that market power finding would have altered the First Amendment analysis.  In particular, Judge Kavanaugh stated that “[i]f the Internet service providers have market power, then the Government may impose open-access or similar carriage obligations.”[44]  “In other words,” Judge Kavanaugh wrote, “if private Internet service providers possess market power, then Turner Broadcasting already gives the Government tools to confront that problem.”[45]  Leaving no doubt on this point, Judge Kavanaugh stated that “[t]he Supreme Court’s First Amendment precedents allow the Government to impose net neutrality obligations on Internet service providers that possess market power.”[46]  

In contrast to the net neutrality case, market power is both present and obvious here.  Indeed, in the Fifth Circuit’s underlying NetChoice decision, the appellate court expressly stated that “each Platform has an effective monopoly over its particular niche of online discourse”[47] and described “[t]he Platforms’ entrenched market power.”[48]  Texas’s own legislative determinations are consistent with these findings of market power, as they reasonably and expressly found that “social media platforms with the largest number of users are common carriers by virtue of their market dominance.”[49]  Therefore, from a First Amendment perspective, the FCC’s 2015 net neutrality decision—with its absence of any market power finding—is entirely different than the NetChoice case.

Second, market considerations readily distinguish dominant social media companies from ISPs.  And the Supreme Court’s First Amendment cases make clear that this distinction matters.  Dominant social media companies exercise market power in a way that is central to public discourse in this country.  As the Supreme Court has indicated, social media companies “provide perhaps the most powerful mechanisms available to a private citizen to make his or her voice heard.”[50]  Therefore, the decision by even a single social media company to ban or censor a speaker can effectively exclude that person from participating in political discourse. 

ISPs are in an entirely different position in this regard.  ISPs are now seeing robust intramodal and intermodal competition.  Wireline, mobile wireless, satellite, and fixed wireless ISPs are all competing today in the broadband market.  Consumers can switch between ISPs and obtain the same product: high-speed Internet service.  This competitive dynamic disciplines the broadband market, as demonstrated by the lack of any evidence that American ISPs are engaged in censorship of the Internet.  Furthermore, due to the consolidation of Internet traffic into a small number of infrastructure providers and applications that ride over the top of broadband networks, and the now-routine encryption of almost all traffic, ISPs are no longer in a position to effectively censor speech, if they ever were.  Your ISP may be able to see when you are using a particular social media platform, but it cannot determine your username, what posts you are viewing, or what content you are sharing, because all of that information is typically encrypted in transit between your device and the platform’s servers, making it illegible to intermediaries like ISPs.  Indeed, ISPs lack the power or mechanism to block a post based on the viewpoints contained in it, notwithstanding a crude, theoretical ability to block entire IP addresses.

Social media platforms are positioned very differently.  As noted above, social media platforms are susceptible to network effects created by a “winner-take-all” dynamic and social graph that tends towards concentration.  A user of one social media platform cannot readily switch to another one and effectively obtain the same type of service.  Content deplatformed from YouTube or Facebook has no comparable medium to reach users in that walled garden or the public square generally; even if posted on an independent blog or video hosting site, it is excluded from the recommendation algorithms populating millions of users’ news feeds, and the advertising revenues that only the largest platforms can deliver.  Put simply, dominant social media companies do not compete in a broader market so much as they are the market.  And unlike ISPs, they can censor content with minute precision.  Thus, the central role that social media platforms play with respect to the public square is utterly unlike the role played by ISPs for First Amendment purposes.

Third, the First Amendment requires the presence and validity of a governmental interest.  If the government fails to marshal evidence of a valid government interest, then any level of First Amendment scrutiny will sweep away the challenged regulation.  The FCC had no valid government interest in promulgating the 2015 net neutrality regulations.  They were, to put it mildly, a solution in search of a problem.  In contrast, as the HB20 litigation demonstrates, the evidence of Silicon Valley’s discriminatory conduct—outlined in the Twitter Files and related revelations—establishes that the government has a sufficiently valid interest, whereas the FCC did not in 2015.  

Fourth, the FCC imposed sweeping and intrusive regulations in 2015 that differ strongly from HB20.  It did not simply adopt net neutrality rules.  Instead, it applied Title II of the Communications Act to ISPs.  Title II imposes a host of regulations and statutory obligations that are entirely distinct from net neutrality. These include government setting of prices, regulation of interconnection and other agreements with competing carriers, and even case-by-case FCC permission to deploy service to new locations.  Additionally, the FCC’s net neutrality rules purported to require ISPs to treat every piece of data the same, subject to reasonable network management.  In this regard, the rules regulated all data that travelled over an ISP’s networks.  Nothing escaped the sweep of that regulation.  In stark contrast, HB20 leaves social media companies substantial leeway in terms of how they handle data and otherwise conduct all aspects of their operations, from advertising sales to back-end management—the law only prohibits social media companies from discriminating based on viewpoints.  It thus sweeps much more narrowly than the FCC’s 2015 Title II regime and presents a different and easier First Amendment case than the one then-Judge Kavanaugh considered.

Fifth, social media censorship involves a form of federal government coercion and state action that was and is completely absent in the ISP context.  Consider Missouri v. Biden or the Congressional testimony by Michael Schellenberger that details what he describes as a “censorship industrial complex” where the government pressured social media companies to censor lawful speech.[51]  The district court’s 155-page findings of fact in Missouri v. Biden describe the Administration’s efforts to control information on social media as “an almost dystopian scenario,” adding that “the United States Government seems to have assumed a role similar to an Orwellian ‘Ministry of Truth.’”[52]  As the case shows, the extent and degree of government officials’ coercive behavior casts doubt on NetChoice’s assertion that HB20 regulates purely private actors.  At a minimum, this government coercion weakens NetChoice’s invocation of private First Amendment rights and further distinguishes this case from attempts to regulate ISPs.  And the difficulty of applying a judicial remedy for this kind of government coercion further cuts in favor of anti-discrimination laws that can prevent it from occurring in the first place.

Sixth, social media companies’ broad assertion of Section 230 rights also makes this case different than the FCC’s 2015 net neutrality rules that then-Judge Kavanaugh found wanting.  The ISPs made no Section 230 claim in that case.  Here, in contrast, NetChoice asserts a Section 230 liability shield against state regulation, and that assertion in turn undermines the strength of their First Amendment claim.  Section 230 provides that a platform is not the “publisher or speaker” of user-generated content.  If NetChoice’s members insist that, under Section 230, they are not the speaker of that content, then their claim to have a particularly strong First Amendment interest in pushing back on the government’s decision to regulate the treatment of that speech would seem to be significantly diminished or extinguished altogether.[53]

For these reasons, the FCC’s 2015 net neutrality regulations are easily distinguished from HB20 for First Amendment purposes.  

IV. Conclusion.

Dominant social media companies play a unique role in the country’s public discourse.  And they now have a clear track record of using their gatekeeper positions to censor or otherwise discriminate against other peoples’ speech.  Despite the serious harms that are flowing from those decisions, Big Tech claims that the First Amendment puts those choices beyond the reach of the government.  The Supreme Court’s First Amendment case law is to the contrary.  

Federal and state governments can craft antidiscrimination laws, in a manner consistent with the First Amendment, that protect and promote a diversity of views.  Far from leading down a slippery slope in either direction, upholding HB20 would preserve both bedrock antidiscrimination protections and viable First Amendment claims, like those then-Judge Kavanaugh identified in the FCC’s 2015 net neutrality case.

Brendan Carr and Nathan Simington are Commissioners at the Federal Communications Commission.  

[1] Packingham v. North Carolina, 137 S. Ct. 1730, 1737 (2017).

[2] See, e.g., Bret Swanson, Covid Censorship Proved to Be Deadly, Wall St. J. (July 7, 2023),

[3] See, e.g., Tex. Bus. & Com. Code § 120.002; Tex. Civ. Prac. & Rem. Code § 143A.002.

[4] See Safeguarding and Securing the Open Internet, WC Docket No. 23-320, Notice of Proposed Rulemaking (Oct. 19, 2023) (dissenting statement of Commissioner Brendan Carr), and Securing the Open Internet, WC Docket No. 23-320, Notice of Proposed Rulemaking (Oct. 19, 2023) (dissenting statement of Commissioner Nathan Simington),  

[5] See also Adam Candeub, Common Carrier Law in the 21st Century, 90 Tenn. L. Rev. 813, 847-56 (2024) (explaining how applying common carriage rules to social media networks could be squared with then-Judge Kavanaugh’s 2017 dissent).

[6] See FCC v. League of Women Voters of California, 468 U.S. 364, 378 (1984) (“Unlike common carriers, broadcasters are entitled under the First Amendment to exercise the widest journalistic freedom consistent with their public duties.”) (cleaned up); Columbia Broad. Sys., Inc. v. Democratic Nat’l Comm., 412 U.S. 94, 106 (1973) (noting that the Communications Act “eliminate[d] the common carrier obligation” for broadcasters to avoid putting them “under the hampering control of being a common carrier and compelled to accept anything and everything that was offered [them] so long as the price was paid” (quoting 67 Cong. Rec. 12,502 (1926))).

[7] Turner Broad. Sys., Inc. v. FCC, 520 U.S. 180, 185 (1997) (Turner II); see also Turner Broad. Sys., Inc. v. FCC, 512 U.S. 622 (1994) (Turner I).

[8] Turner I, 512 U.S. at 661; see also 47 U. S. C. § 531(b) (requiring cable systems to set aside channels for public access).  We note that the market factors considered in Turner related to the broadcast and cable television markets as they existed in the 1990s—a point in time when Congress viewed cable as a central means for Americans to obtain information about the important social and political issues of the day.  Regardless of how an analysis of those broadcast and cable market factors play out if they were analyzed under the First Amendment today, the regulation of large social media platforms is a straightforward case given their ongoing market dominance, which rivals or exceeds the position held by cable in the 1990s.

[9] Pruneyard Shopping Ctr. v. Robins, 447 U.S. 74 (1980).

[10] Id. at 87-88.

[11] Zauderer v. Office of Disc. Counsel, 471 U.S. 626, 651 (1985).

[12] Miami Herald Pub. Co., Div. of Knight Newspapers, Inc. v. Tornillo, 418 U.S. 241 (1974).

[13] Pac. Gas & Elec. Co. v. Public Utilities Comm’n of Cal., 475 U.S. 1 (1986).

[14] Hurley v. Irish-American Gay, Lesbian Bisexual Group, 515 U.S. 557 (1995).

[15] Turner I, 512 U.S. at 656.

[16] Id. 

[17] 47 U.S.C. § 541(c)

[18] Turner I, 512 U.S. at 656.

[19] Id.

[20] Id.

[21] Id. at 657. 

[22] Hurley, 515 U.S. at 577.

[23] Packingham, 137 S. Ct. at 1737.

[24] Mark Jurkowitz and Jeffrey Gottfried, Twitter is the go-to social media site for U.S. journalists, but not for the public, Pew Res. Ctr. (June 27, 2022),

[25] Turner I, 512 U.S. at 655-56.

[26] Tornillo, 418 U.S. at 258.

[27] Id.

[28] Turner I, 512 U.S. at 636-37.

[29] 47 U.S.C. §§ 230 (a)(3), (c)(1).

[30] Tornillo, 418 U.S. at 248, 261 (White, J., concurring).

[31] Id. at 244.

[32] Turner I, 512 U.S. at 629.

[33] Id. at 655-56.

[34] See Rumsfeld v. Forum for Academic and Institutional Rights, Inc., 547 U.S. 47, 65 (2006) (“Nothing about [hosting military] recruiting suggests that law schools agree with any speech by recruiters, and nothing in the Solomon Amendment restricts what the law schools may say about the military’s policies.”). 

[35] Turner I, 512 U.S. at 642.

[36] Id.

[37] Id. (noting that content-neutral laws greatly reduce the risk “of excising certain ideas or viewpoints from the public dialogue”); see also Ward v. Rock Against Racism, 491 U.S. 781, 791 (1989) (noting that a critical factor is “whether the government has adopted a regulation of speech because of [agreement or] disagreement with the message it conveys”).

[38] Turner I, 512 U.S. at 645.

[39] Forum for Academic and Institutional Rights, Inc., 547 U.S. at 62.

[40] Turner I, 512 U.S. at 653-54 (discussing Tornillo, 418 U.S. at 256-58).

[41] Pac. Gas & Elec., 475 U.S. at 8 (“In appearance no different from a small newspaper, Progress’ contents range from energy-saving tips to stories about wildlife conservation, and from billing information to recipes”).

[42] Turner I, 512 U.S. at 643; Forum for Academic and Institutional Rights, Inc., 547 U.S. at 64.

[43] Protecting and Promoting the Open Internet, Report and Order, 30 FCC Red. 5601 (2015). Nor does upholding HB20 compel a determination that the FCC’s long repealed Fairness Doctrine would survive First Amendment scrutiny today.  Among other things, HB20 operates in a fundamentally different way than the Fairness Doctrine did, rendering any attempted analogy inapt.  The Fairness Doctrine’s obligations were triggered by the content of broadcasters’ speech, and they imposed an affirmative obligation on broadcasters to air certain content as a result.  That is not at all how HB20 operates—it is not triggered by the content of any speech, and it merely prohibits social media companies from engaging in viewpoint discrimination.

[44] U.S. Telecom Ass’n v. FCC, 855 F.3d 381, 434-435 (D.C. Cir. 2017) (en banc) (Kavanaugh, J., dissenting).

[45] Id. at 435.                                                            

[46] Id. (emphasis in original).

[47] NetChoice, LLC v. Paxton, 49 F.4th 439, 476 (5th Cir. 2022).

[48] Id.

[49] See Bill Report, C.S.H.B. 20,

[50] Packingham, 137 S. Ct. at 1737.

[51] See Testimony by Michael Shellenberger to The House Select Committee on the Weaponization of the Federal Government (Mar. 9, 2023),  

[52] Missouri v. Biden, _ F. Supp. 3d _, 2023 WL 4335270 (W.D. La., July 4, 2023), available at

[53] See Candeub, supra note 5, at 861-62 (arguing that “liability rules undoubtedly create context and expectation for whether speech is the expression of any particular speaker”).

Print Friendly, PDF & Email