Free Speech and the Regulation of Social Media Content

Free Speech and the Regulation of

Social Media Content

Valerie C. Brannon

Legislative Attorney

March 27, 2019

Congressional Research Service

7-....

www.crs.gov

R45650

SUMMARY

Free Speech and the Regulation of

Social Media Content

As the Supreme Court has recognized, social media sites like Facebook and Twitter have become

important venues for users to exercise free speech rights protected under the First Amendment.

Commentators and legislators, however, have questioned whether these social media platforms

are living up to their reputation as digital public forums. Some have expressed concern that these

sites are not doing enough to counter violent or false speech. At the same time, many argue that

the platforms are unfairly banning and restricting access to potentially valuable speech.

R45650

March 27, 2019

Valerie C. Brannon

Legislative Attorney

-re-acte--@crs.loc.gov

For a copy of the full report,

please call 7-.... or visit

www.crs.gov.

Currently, federal law does not offer much recourse for social media users who seek to challenge a social media provider’s

decision about whether and how to present a user’s content. Lawsuits predicated on these sites’ decisions to host or remove

content have been largely unsuccessful, facing at least two significant barriers under existing federal law. First, while

individuals have sometimes alleged that these companies violated their free speech rights by discriminating against users’

content, courts have held that the First Amendment, which provides protection against state action, is not implicated by the

actions of these private companies. Second, courts have concluded that many non-constitutional claims are barred by Section

230 of the Communications Decency Act, 47 U.S.C. § 230, which provides immunity to providers of interactive computer

services, including social media providers, both for certain decisions to host content created by others and for actions taken

“voluntarily” and “in good faith” to restrict access to “objectionable” material.

Some have argued that Congress should step in to regulate social media sites. Government action regulating internet content

would constitute state action that may implicate the First Amendment. In particular, social media providers may argue that

government regulations impermissibly infringe on the providers’ own constitutional free speech rights. Legal commentators

have argued that when social media platforms decide whether and how to post users’ content, these publication decisions are

themselves protected under the First Amendment. There are few court decisions evaluating whether a social media site, by

virtue of publishing, organizing, or even editing protected speech, is itself exercising free speech rights. Consequently,

commentators have largely analyzed the question of whether the First Amendment protects a social media site’s publication

decisions by analogy to other types of First Amendment cases. There are at least three possible frameworks for analyzing

governmental restrictions on social media sites’ ability to moderate user content.

First, using the analogue of the company town, social media sites could be treated as state actors who are themselves bound

to follow the First Amendment when they regulate protected speech. If social media sites were treated as state actors under

the First Amendment, then the Constitution itself would constrain their conduct, even absent legislative regulation. The

second possible framework would view social media sites as analogous to special industries like common carriers or

broadcast media. The Court has historically allowed greater regulation of these industries’ speech, given the need to protect

public access for users of their services. Under the second framework, if special aspects of social media sites threaten the use

of the medium for communicative or expressive purposes, courts might approve of content-neutral regulations intended to

solve those problems. The third analogy would treat social media sites like news editors, who generally receive the full

protections of the First Amendment when making editorial decisions. If social media sites were considered to be equivalent

to newspaper editors when they make decisions about whether and how to present users’ content, then those editorial

decisions would receive the broadest protections under the First Amendment. Any government regulations that alter the

editorial choices of social media sites by forcing them to host content that they would not otherwise transmit, or requiring

them to take down content they would like to host, could be subject to strict scrutiny. A number of federal trial courts have

held that search engines exercise editorial judgment protected by the First Amendment when they make decisions about

whether and how to present specific websites or advertisements in search results, seemingly adopting this last framework.

Which of these three frameworks applies will depend largely on the particular action being regulated. Under existing law,

social media platforms may be more likely to receive First Amendment protection when they exercise more editorial

discretion in presenting user-generated content, rather than if they neutrally transmit all such content. In addition, certain

types of speech receive less protection under the First Amendment. Courts may be more likely to uphold regulations targeting

certain disfavored categories of speech such as obscenity or speech inciting violence. Finally, if a law targets a social media

site’s conduct rather than speech, it may not trigger the protections of the First Amendment at all.

Congressional Research Service

Free Speech and the Regulation of Social Media Content

Contents

Existing Legal Barriers to Private Lawsuits Against Social Media Providers ................................ 4

First Amendment: State Action Requirement............................................................................ 5

Section 230 of the CDA ............................................................................................................ 9

Section 230(c)(1) ............................................................................................................... 11

Section 230(c)(2) .............................................................................................................. 13

First Amendment Limits on Government Regulation of Social Media Content ........................... 15

Background Principles: First Amendment Protections Online................................................ 17

Social Media Sites: Providing a Digital Public Square ........................................................... 21

Social Media Sites as Company Towns ............................................................................ 23

Social Media Sites as Broadcasters or Cable Providers .................................................... 27

Social Media Sites as Editors ............................................................................................ 33

Considerations for Congress.......................................................................................................... 40

Contacts

Author Contact Information .......................................................................................................... 43

Congressional Research Service

Free Speech and the Regulation of Social Media Content

ne of the core purposes of the First Amendment’s Free Speech Clause is to foster “an

uninhibited marketplace of ideas,”1 testing the “truth” of various ideas “in the

competition of the market.”2 Social media sites3 provide one avenue for the

transmission of those ideas.4 The Supreme Court has recognized that the internet in

general, and social media sites in particular, are “important places” for people to

“speak and listen,” observing that “social media users employ these websites to engage in a wide

array of protected First Amendment activity.”5 Users of social media sites such as Facebook,

Twitter, YouTube, or Instagram can use these platforms to post art or news,6 debate political

issues,7 and document their lives.8 In a study conducted in early 2018, the Pew Research Center

found that 68% of U.S. adults use Facebook, 35% use Instagram, and 24% report using Twitter.9

These sites not only allow users to post content, they also connect users with each other, allowing

users to seek out friends and content and often recommending new connections to the user.10 On

most social media platforms, users can then send content to specific people, or set permissions

allowing only certain people to view that content. Through human curation and the use of

algorithms, these platforms decide how content is displayed to other users.11 In curating this

content, social media sites may also edit user content, combine it, or draft their own additions to

that content.12 These platforms are generally free to users, and make revenue by selling targeted

O

1 Virginia v. Hicks, 539 U.S. 113, 119 (2003).

2 Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting). Accord, e.g., Red Lion Broad. Co. v.

FCC, 395 U.S. 367, 390 (1969).

3 The report does not precisely define the term “social media site,” given that most of the cases defining First

Amendment rights on the internet focus more on various companies’ actions than their character. For one possible

taxonomy of internet “intermediaries” proposed for future First Amendment analysis, see David S. Ardia, Free Speech

Savior or Shield for Scoundrels: An Empirical Study of Intermediary Immunity under Section 230 of the

Communications Decency Act, 43 LOY. L.A. L. REV. 373, 386–87 (2010).

4 See, e.g., Nicholas A. Primrose, Has Society Become Tolerant of Further Infringement on First Amendment Rights?,

19 BARRY L. REV. 313, 333 (2014) (“Social media is the ideal ‘marketplace’ for the 21st century; it creates a dynamic

place for every conceivable opinion to be expressed and shared.”).

5 Packingham v. North Carolina, 137 S. Ct. 1730, 1735–36 (2017).

6 See, e.g., Elisa Shearer, Social Media Outpaces Print Newspapers in the U.S. as a News Source, PEW RESEARCH CTR.

FACT TANK (Dec. 10, 2018), https://pewrsr.ch/2UvWPSe.

7 See, e.g., Kristen Bialik, 14% of Americans Have Changed their Mind about an Issue Because of Something They Saw

on Social Media, PEW RESEARCH CTR. FACT TANK (Aug. 15, 2018), https://pewrsr.ch/2Bc5TWF.

8 See, e.g., Aaron Smith & Monica Anderson, Social Media Use in 2018, PEW RESEARCH CTR. FACT TANK (Mar. 1,

2018), https://pewrsr.ch/2FDfiFd.

9 Id.

10 While many social media platforms are primarily centered around connections with friends or followers, this is not

universally true. For example, commentators have discussed the app TikTok as an alternative to this friend-centric

approach, noting that the platform opens to a page full of content that an algorithm has determined may interest the user

based on past usage habits, rather than to a feed of friends’ content. See, e.g., Caroline Haskins, TikTok Can’t Save Us

from Algorithmic Content Hell, VICE MOTHERBOARD (Jan. 31, 2019), https://bit.ly/2Jlu1dF; John Herrman, How TikTok

Is Rewriting the World, N.Y. TIMES (Mar. 10, 2019), https://nyti.ms/2Hm0ygS.

11 See, e.g., Paul Hitlin & Lee Rainie, Facebook Algorithms and Personal Data, PEW RESEARCH CTR. (Jan. 16, 2019),

https://pewrsr.ch/2Hnqr1o; Will Oremus, Twitter’s New Order, SLATE (Mar. 5, 2017), https://bit.ly/2lMs0pU.

12 For example, Twitter sometimes creates “Moments,” which it describes as “curated stories showcasing the very best

of what’s happening on Twitter.” About Moments, TWITTER, https://bit.ly/2x0AxOt (last visited Mar. 27, 2019).

Congressional Research Service

1

Free Speech and the Regulation of Social Media Content

advertising space,13 among other things.14 Thus, social media sites engage in a wide variety of

activities, at least some of which entail hosting—and creating—constitutionally protected speech.

Social media companies have recognized their role in providing platforms for speech.15 To take

one example, in a September 2018 hearing before the Senate Select Committee on Intelligence,

the founder and Chief Executive Officer of Twitter, Jack Dorsey, repeatedly referred to Twitter as

a “digital public square,” emphasizing the importance of “free and open exchange” on the

platform.16 Critically, however, social media sites also have content-moderation policies under

which they may remove certain content. Further, these sites determine how content is presented:

who sees it, when, and where. As one scholar has said, social media sites “create rules and

systems to curate speech out of a sense of corporate social responsibility, but also . . . because

their economic viability depends on meeting users’ speech and community norms.”17 Speech

posted on the internet “exists in an architecture of privately owned websites, servers, routers, and

backbones,” and its existence online is subject to the rules of those private companies.18

Consequently, one First Amendment scholar predicted ten years ago that “the most important

decisions affecting the future of freedom of speech will not occur in constitutional law; they will

be decisions about technological design, legislative and administrative regulations, the formation

of new business models, and the collective activities of end-users.”19

Social media companies have come under increased scrutiny regarding the type of user content

that they allow to be posted on their sites, and the ways in which they may promote—or

deemphasize—certain content. A wide variety of people have expressed concern that these sites

do not do enough to counter harmful, offensive, or false content.20 At the same time, others have

13 See, e.g., Kalev Leetaru, What Does It Mean For Social Media Platforms To “Sell” Our Data?, FORBES (Dec. 15,

2018), https://bit.ly/2W5MfRL; Louise Matsakis, Facebook’s Targeted Ads Are More Complex Than It Lets On,

WIRED (Apr. 25, 2018, 4:04 PM), https://bit.ly/2DZ1CG0.

14

Some otherwise free sites include a subscription option, allowing certain users to pay fees to access premium content

or unlock certain features. See, e.g., LinkedIn Free Accounts and Premium Subscriptions, LINKEDIN,

https://bit.ly/2sRWDhR (last visited Mar. 27, 2019).

15 See, e.g., Foreign Influence Operations’ Use of Social Media Platforms: Hearing Before the S. Select Comm. on

Intelligence, 115th Cong. (Sept. 5, 2018) [hereinafter Hearing on Foreign Influence Operations] (statement of Jack

Dorsey, CEO of Twitter) (“[W]e believe that the people use Twitter as they would a public square and they often have

the same expectations that they would have of any public space. For our part, we see our platform as hosting and

serving conversations.”); Facebook, Social Media Privacy, and the Use and Abuse of Data: Hearing Before the S.

Comm. on the Judiciary and the S. Comm. on Commerce, Sci. & Transp., 115th Cong. (Apr. 10, 2018) (statement of

Mark Zuckerberg, CEO of Facebook) (“[W]e consider ourselves to be a platform for all ideas.”).

16 Hearing on Foreign Influence Operations, supra note 15.

17 Kate Klonick, The New Governors: The People, Rules, and Processes Governing Online Speech, 131 HARV. L. REV.

1598, 1625 (2018).

18 Jonathan Peters, The “Sovereigns of Cyberspace” and State Action: The First Amendment’s Application—or Lack

Thereof—to Third-Party Platforms, 32 BERKELEY TECH. L.J. 989, 990 (2017).

19 Jack M. Balkin, Free Speech and Press in the Digital Age: The Future of Free Expression in a Digital Age, 36 PEPP.

L. REV. 427, 427 (2009).

20 E.g., Anne Applebaum, Regulate Social Media Now. The Future of Democracy is at Stake., WASH. POST (Feb. 1,

2019), https://wapo.st/2T4SmYS; John Carroll & David Karpf, How Can Social Media Firms Tackle Hate Speech?,

KNOWLEDGE AT WHARTON U. PENN. (Sept. 22, 2018), https://whr.tn/2DuFeFH; David Dayen, Ban Targeted

Advertising, THE NEW REPUBLIC (Apr. 10, 2018), https://bit.ly/2GRrT7z; Danielle Kurtzleben, Did Fake News On

Facebook Help Elect Trump? Here’s What We Know, NPR (Apr. 11, 218, 7:00 AM), https://n.pr/2GPdMDP; Caroline

O’Donovan & Logan McDonald, YouTube Continues To Promote Anti-Vax Videos As Facebook Prepares To Fight

Medical Misinformation, BUZZFEED NEWS (Feb. 20, 2019), https://bit.ly/2IWjIN5. Cf., e.g., Mehreen Khan, More

‘Hate Speech’ Being Removed from Social Media, FINANCIAL TIMES (Feb. 2, 2019), https://on.ft.com/2EO6YCR.

Congressional Research Service

2

Free Speech and the Regulation of Social Media Content

argued that the platforms take down or deemphasize too much legitimate content.21 In the

September 2018 hearing referenced above, Sheryl Sandberg, the Chief Operating Officer of

Facebook, expressed the difficulty of determining what types of speech would violate company

standards barring hate speech.22 Both Dorsey and Facebook founder and Chief Executive Officer

Mark Zuckerberg have been asked to respond to allegations of political bias in their platforms’

content moderation decisions at hearings before House and Senate committees.23 Commentators

and legislators alike have questioned whether social media sites’ content policies are living up to

the free speech ideals they have espoused.24 As a result, some, including Members of Congress,25

have called for regulation of social media platforms, focused on the way those companies police

content.26

In light of this public policy debate, this report begins by outlining the current legal framework

governing social media sites’ treatment of users’ content, focusing on the First Amendment and

Section 230 of the Communications Decency Act of 1996 (CDA).27 As explained below, under

existing law, lawsuits predicated on these sites’ decisions to remove or to host content have been

largely unsuccessful because of (1) doctrines that prevent the First Amendment from being

applied to private social media companies,28 and (2) Section 230 of the CDA, which often

protects social media companies from being held liable under federal or state laws for these

21 See, e.g., Mike Isaac & Sydney Ember, Live Footage of Shootings Forces Facebook to Confront New Role, N.Y.

TIMES (July 8, 2016), https://nyti.ms/2CpLI37; Eli Rosenberg, Facebook Censored a Post for ‘Hate Speech.’ It Was the

Declaration of Independence., WASH. POST. (July 5, 2018), https://wapo.st/2SRj4zN; Liam Stack, What Is a ‘Shadow

Ban,’ and Is Twitter Doing It to Republican Accounts?, N.Y. TIMES (July 26, 2018), https://nyti.ms/2NRWLYC; Jillian

C. York & Karen Gullo, Offline/Online Project Highlights How the Oppression Marginalized Communities Face in the

Real World Follows Them Online, ELEC. FRONTIER FOUND. (Mar. 6, 2018), https://bit.ly/2ulHdon.

22 Hearing on Foreign Influence Operations, supra note 15. See also, e.g., Jason Koebler & Joseph Cox, The

Impossible Job: Inside Facebook’s Struggle to Moderate Two Billion People, VICE MOTHERBOARD (Aug. 23, 2018,

1:15 PM), https://bit.ly/2wtRboq.

23 Hearing on Foreign Influence Operations, supra note 15; Twitter: Transparency and Accountability: Hearing Before

the H. Comm. on Energy & Commerce, 115th Cong. (Sept. 5, 2018); Facebook, Social Media Privacy, and the Use and

Abuse of Data: Hearing Before the S. Comm. on the Judiciary and the S. Comm. on Commerce, Sci. & Transp., 115th

Cong. (Apr. 10, 2018). See also, e.g., Langdon v. Google, Inc., 474 F. Supp. 2d 622, 627 (D. Del. 2007) (“Plaintiff

alleges that [Google’s] rejection or acceptance of ads is based upon . . . the political viewpoint of the ad . . . .”).

24 See, e.g., William Cummings, Republican Lawmakers Go after Facebook CEO Zuckerberg for Anti-Conservative

Bias, USA TODAY (Apr. 11, 2018), https://bit.ly/2BTQ5VH; Mattathias Schwartz, Facebook and Twitter’s Rehearsed

Dance on Capitol Hill, THE NEW YORKER (Sept. 6, 2018), https://bit.ly/2VqgDXG.

25 To take one example, a bill was introduced in the 115th Congress, the Honest Ads Act, that would have ensured that

disclosure requirements for political advertisements apply to advertising on social media. S. 1989, 115th Cong. (2017).

Both Facebook and Twitter expressed support for this act. E.g., Selina Wang, Twitter Follows Facebook in Endorsing

Senate’s ‘Honest Ads’ Act, BLOOMBERG (Apr. 10, 2018), https://bloom.bg/2Y8m7rm.

26 See, e.g., Mark Epstein, Google’s Effort To Undermine Free Speech Strengthens Case for Regulating Big Tech, THE

HILL (Aug. 31, 2017), https://bit.ly/2Rs0OjT; Makena Kelly, Sen. Josh Hawley Is Making the Conservative Case

Against Facebook, THE VERGE (Mar. 19, 2019, 8:00 AM), https://bit.ly/2Tf9S8B; David McCabe, Key Republican

Warns Big Tech: Step Up or Be Regulated, AXIOS (Feb. 28, 2018), https://bit.ly/2OlgqC5; Katy Steinmetz, Lawmakers

Hint at Regulating Social Media During Hearing with Facebook and Twitter Execs, TIME (Sept. 5, 2018),

https://bit.ly/2Rv4tgX; Craig Timberg, Tony Romm & Elizabeth Dwoskin, Lawmakers Agree Social Media Needs

Regulation, But Say Prompt Federal Action Is Unlikely, WASH. POST (Apr. 11, 2018), https://wapo.st/2OekLqD. Cf.,

e.g., Ben Brody, Ted Cruz Echoes Elizabeth Warren’s Criticism of Facebook’s Power, BLOOMBERG (Mar. 12, 2019,

3:25 PM), https://bloom.bg/2JAfqLF; Sen. Elizabeth Warren, Here’s How We Can Break Up Big Tech, MEDIUM (Mar.

8, 2019), https://bit.ly/2SRNIt5.

27 47 U.S.C. § 230. While this provision is often referred to as Section 230 of the CDA, it was enacted as Section 509

of the Telecommunications Act of 1996, which amended Section 230 of the Telecommunications Act of 1934. Pub. L.

No. 104-104, 110 Stat. 137 (1996).

28 See, e.g., Peters, supra note 18, at 992.

Congressional Research Service

3

Free Speech and the Regulation of Social Media Content

decisions.29 The debate over whether the federal government should fill this legal vacuum has

raised the question as to whether and to what extent the federal government can regulate the way

social media sites present users’ content, either to require these sites to take down, restrict access

to, or qualify certain types of content, or, on the other hand, protect users’ rights to post content

on those sites.30 Such government regulation would constitute state action that implicates the First

Amendment.31 While the issue largely remains an open question in the courts, the First

Amendment may provide some protection for social media companies when they make content

presentation decisions, limiting the federal government’s ability to regulate those decisions.32 The

extent of any free speech protections will depend on how courts view social media companies and

the specific action being regulated.33

Accordingly, the bulk of this report explores how the First Amendment applies to social media

providers’ content presentation decisions. Looking to three possible analogues drawn from

existing First Amendment law, the report explores whether social media companies could be

viewed in the same way as company towns, broadcasters, or newspaper editors.34 The report also

explains the possible regulatory implications of each First Amendment framework as Congress

considers the novel legal issues raised by the regulation of social media.

Existing Legal Barriers to Private Lawsuits Against

Social Media Providers

Under current federal law, social media users may face at least two significant barriers if they

attempt to sue a social media provider for its decisions about hosting or limiting access to users’

content. The first, which likely applies only to lawsuits predicated on a platform’s decision to

remove rather than allow content,35 is the state action requirement of the First Amendment. The

state action doctrine provides that constitutional free speech protections generally apply only

29 See, e.g., Eric Goldman, Online User Account Termination and 47 U.S.C. § 230(c)(2), 2 U.C. IRVINE L. REV. 659,

662 (2012).

30 See supra notes 20 and 21.

31 Cf., e.g., Reno v. ACLU, 521 U.S. 844, 879 (1997) (holding that provisions of the CDA prohibiting the transmission

of certain “indecent” or “patently offensive” material to children are unconstitutional under the First Amendment).

32 See, e.g., Kate Klonick, The New Governors: The People, Rules, and Processes Governing Online Speech, 131

HARV. L. REV. 1598, 1612–13 (2018).

33 See, e.g., U.S. Telecom Ass’n v. FCC, 825 F.3d 674, 741–42 (D.C. Cir. 2016) (holding that net neutrality rules do

not violate the First Amendment, as applied to “the provision of internet access as common carriage,” but noting that

“insofar as a broadband provider might offer its own content—such as a news or weather site—separate from its

internet access service, the provider would receive the same protection under the First Amendment as other producers

of internet content”). In particular, legal commentators have disagreed regarding whether content originally created by

users should be treated in some circumstances as the platforms’ speech, given that these platforms often make editorial

decisions about whether and how to present that user-generated content. See, e.g., Stuart Minor Benjamin, Determining

What “The Freedom of Speech” Encompasses, 60 DUKE L.J. 1673, 1680 (2011).

34 See, e.g., Klonick, supra note 32, at 1658.

35 As discussed infra, “First Amendment: State Action Requirement,” the First Amendment protects against

government actions “abridging the freedom of speech.” U.S. CONST. amend. I. Logically, while parties may abridge

speech when they take down or restrict access to user content, they likely do not abridge speech by disseminating that

speech. The Supreme Court has held that the government can violate the First Amendment when, for example, it

requires speakers to accompany their speech with disclaimers or other notices. E.g., Nat’l Inst. of Family & Life

Advocates v. Becerra, 138 S. Ct. 2361, 2378 (2018). Nonetheless, in this context, the Court was not concerned by the

state’s decision to allow the original content, but by the fact that it required additional content, and was particularly

concerned that this additional content may “drown[] out” the original speech. See id. at 2378.

Congressional Research Service

4

Free Speech and the Regulation of Social Media Content

when a person is harmed by an action of the government, rather than a private party.36 The second

legal barrier is the CDA’s Section 230, which offers broad immunity to “interactive computer

service” providers.37 Section 230(c)(1) provides immunity from any lawsuit that seeks to hold a

service provider liable for publishing information that was created by an “information content

provider,” effectively protecting social media sites from liability for hosting content.38 By

contrast, Section 230(c)(2) provides immunity for sites that take good faith action to restrict

access to content that the provider or users deem “obscene, lewd, lascivious, filthy, excessively

violent, harassing, or otherwise objectionable.”39 Thus, federal law does not currently provide a

recourse for many users who would like to challenge a social media site’s decision to ban or

restrict content, or to host content—and may affirmatively bar liability in certain circumstances.

First Amendment: State Action Requirement

The Free Speech Clause of the First Amendment provides that “Congress shall make no law . . .

abridging the freedom of speech”40 and applies to the “State[s]” through the Fourteenth

Amendment.41 Thus, the First Amendment, like other constitutional guarantees, generally applies

only against government action.42 As the Supreme Court has said, “while statutory or common

law may in some situations extend protection or provide redress against a private corporation or

person who seeks to abridge the free expression of others, no such protection or redress is

provided by the Constitution itself.”43 However, the Supreme Court has, in limited circumstances,

allowed First Amendment claims to proceed against seemingly private parties that abridge

protected speech.44

The clearest example of the Court extending the First Amendment to apply to the actions of a

private party comes from Marsh v. Alabama, where the Court held that the First Amendment

prohibited the punishment of a resident of a company-owned town for distributing religious

literature.45 While the town in question was owned by a private corporation, “it ha[d] all the

characteristics of any other American town,” including residences, businesses, streets, utilities,

public safety officers, and a post office.46 Under these circumstances, the Court held that “the

36 See, e.g., Lloyd Corp. v. Tanner, 407 U.S. 551, 567 (1972).

37 47 U.S.C. § 230(c). See, e.g., Klayman v. Zuckerberg, 753 F.3d 1354, 1358–59 (D.C. Cir. 2014) (discussing scope of

immunity provided by Section 230).

38 47 U.S.C. § 230(c)(1).

39 Id. § 230(c)(2).

40 U.S. CONST. amend. I (emphasis added). Although the text of the First Amendment refers to “Congress,” it has long

been understood to restrict action by the executive branch as well. See, e.g., Columbia Broad. Sys., Inc. v. Democratic

Nat’l Comm., 412 U.S. 94, 160 (1973) (Douglas, J., concurring) (describing First Amendment as restricting Congress,

whether “acting directly or through any of its agencies such as the FCC”); see generally, e.g., Daniel J. Hemel,

Executive Action and the First Amendment’s First Word, 40 PEPP. L. REV. 601 (2013).

41 44 Liquormart v. Rhode Island, 517 U.S. 484, 489 n.1 (1996); U.S. CONST. amend. XIV (“[N]o State shall . . .

deprive any person of life, liberty, or property, without due process of law . . . .”).

42

See, e.g., Lloyd Corp. v. Tanner, 407 U.S. 551, 567 (1972).

43 Hudgens v. NLRB, 424 U.S. 507, 513 (1976).

44 See, e.g., Blum v. Yaretsky, 457 U.S. 991, 1004 (1982); Marsh v. Alabama, 326 U.S. 501, 508–09 (1946).

45 Marsh, 326 U.S. at 509. A state statute “ma[de] it a crime to enter or remain on the premises of another after having

been warned not to do so”; the resident had been warned that, pursuant to a company policy, she could not distribute

religious literature without a permit, and she subsequently disregarded that warning and refused to leave a sidewalk. Id.

at 503–04. Accordingly, although the case involved a criminal prosecution brought by the State of Alabama, liability

turned on the town’s ability to prevent residents from distributing literature without a permit. See id.

46 Id. at 502–03.

Congressional Research Service

5

Free Speech and the Regulation of Social Media Content

corporation’s property interests” did not “settle the question”47: “[w]hether a corporation or a

municipality owns or possesses the town[,] the public in either case has an identical interest in the

functioning of the community in such manner that the channels of communication remain free.”48

Consequently, the corporation could not be permitted “to govern a community of citizens” in a

way that “restrict[ed] their fundamental liberties.”49 The Supreme Court has described Marsh as

embodying a “public function” test, under which the First Amendment will apply if a private

entity exercises “powers traditionally exclusively reserved to the State.”50

Since Marsh was issued in 1946, however, it has largely been limited to the facts presented in that

case. The Supreme Court extended the Marsh decision in 1968: in Amalgamated Food Employees

Union v. Logan Valley Plaza, the Court held that a private shopping mall could not prevent

individuals from peacefully picketing on the premises, noting similarities between “the business

block in Marsh and the shopping center” at issue in that case.51 However, the Court subsequently

disclaimed Logan Valley in Hudgens v. NLRB, rejecting the idea that “large self-contained

shopping center[s]” are “the functional equivalent of a municipality.”52 Instead, the Court held

that in Hudgens, where a shopping center manager had threatened to arrest picketers for

trespassing, “the constitutional guarantee of free expression ha[d] no part to play.”53 As a result,

the picketers “did not have a First Amendment right to enter this shopping center for the purpose

of advertising their strike.”54 In another decision in which the Supreme Court held that the First

Amendment did not prevent a shopping center from banning the distribution of handbills, the

Court distinguished Marsh by noting that “the owner of the company town was performing the

full spectrum of municipal powers and stood in the shoes of the State.”55 By contrast, the disputed

shopping center had not assumed “municipal functions or power.”56 The fact that the shopping

center was generally open to the public did not qualify as a “dedication of [the] privately owned

and operated shopping center to public use” sufficient “to entitle respondents to exercise therein

the asserted First Amendment rights.”57

Apart from the factual circumstances presented by the company town that exercises powers

“traditionally” and “exclusively” held by the government,58 the Court has sometimes applied the

First Amendment against private parties if they have a “sufficiently close relationship” to the

government.59 Such circumstances may exist where a private company “is subject to extensive

47 Id. at 505.

48 Id. at 507. See also id. at 508 (noting that residents of company towns, like residents of other towns, “must make

decisions which affect the welfare of community and nation,” and that to do this, they must have access to

“uncensored” information).

49 Id. at 509.

50 Jackson v. Metro. Edison Co., 419 U.S. 345, 352 (1974) (emphasis added). Accord Lugar v. Edmonson Oil Co., 457

U.S. 922, 939 (1982); Flagg Bros., Inc. v. Brooks, 436 U.S. 149, 158–59 (1978).

51 391 U.S. 308, 317 (1968). In dissent, Justice Black would have ruled that the picketers could not, “under the guise of

exercising First Amendment rights, trespass on . . . private property for the purpose of picketing.” Id. at 329 (Black, J.,

dissenting).

52 Hudgens v. NLRB, 424 U.S. 507, 520 (1976).

53 Id. at 521.

54 Id.

55 Lloyd Corp. v. Tanner, 407 U.S. 551, 569 (1972).

56 Id.

57 Id. at 569–70.

58 Jackson v. Metro. Edison Co., 419 U.S. 345, 352 (1974).

59 See Pub. Utils. Comm’n v. Pollak, 343 U.S. 451, 462 (1952) (holding that such a relationship existed where the

private company operated a public utility that represented a “substantial monopoly” under congressional authority and,

Congressional Research Service

6

Free Speech and the Regulation of Social Media Content

state regulation”—although government regulation alone is not sufficient to establish the state

action requirement.60 Instead, the inquiry in such a case is “whether there is a sufficiently close

nexus between the State and the challenged action of the regulated entity so that the action of the

latter may be fairly treated as that of the State itself.”61 In a 2001 case, the Supreme Court held

that a state athletic association, while “nominally private,” should be subject to First Amendment

standards because of “the pervasive entwinement of public institutions and public officials in its

composition and workings.”62

Some plaintiffs have argued that various internet companies, including some social media sites,

should be treated as state actors subject to the First Amendment when those companies take down

or restrict access to their speech. Courts have rejected these claims.63 Many of these decisions

have involved relatively terse applications of existing Supreme Court precedent.64 In a few cases,

however, federal district courts have explored the application of these state action cases in more

detail.

First, lower courts have repeatedly held that social media sites do not meet the “exclusive public

function test”65 and are not akin to a company town.66 In so holding, courts have recognized that,

under prevailing Supreme Court case law, private actors are not “state actors subject to First

Amendment scrutiny merely because they hold out and operate their private property as a forum

for expression of diverse points of view.”67 Accordingly, they have held that the mere fact that

social media providers hold their networks open for use by the public is insufficient to make them

more importantly, the company operated “under the regulatory supervision” of a governmental agency, and the

particular action being challenged involved action by that agency).

60 Jackson v. Metro. Edison Co., 419 U.S. 345, 350 (1974); see also Brentwood Acad. v. Tenn. Secondary Sch. Athletic

Ass’n, 531 U.S. 288, 295 (2001).

61 Jackson, 419 U.S. at 351. In Burton v. Wilmington Parking Authority, the Supreme Court held that a private parking

authority had to comply with the Equal Protection Clause of the Fourteenth Amendment where a garage that it operated

was “erected and maintained with public funds by an agency of the State to serve a public purpose.” 365 U.S. 715, 724

(1961). In that case, the Court believed that the state had “so far insinuated itself into a position of interdependence

with [the private party] that it must be recognized as a joint participant in the challenged activity.” Id. at 725. Courts

have sometimes described Burton as creating a “symbiotic relationship” test for state action, separate from the “nexus”

standard outlined in Jackson. See, e.g., Jackson, 419 U.S. at 357; Perkins v. Londonderry Basketball Club, 196 F.3d 13,

20 (1st Cir. 1999); Gallagher v. Neil Young Freedom Concert, 49 F.3d 1442, 1447 (10th Cir. 1995); Fitzgerald v.

Mountain Laurel Racing Inc., 607 F.2d 589, 595 (3d Cir. 1979). Cf., e.g., Marie v. Am. Red Cross, 771 F.3d 344, 362

(6th Cir. 2014) (describing “the symbiotic relationship or nexus test” as one of four possible state action inquiries).

62 Brentwood Acad., 531 U.S. at 298.

63 E.g., Jayne v. Google Internet Search Engine Founders, 63 Fed. Appx. 268, 268 (3d Cir. 2008) (per curiam); Green v.

Am. Online (AOL), 318 F.3d 465, 472 (3rd Cir. 2003); Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018

U.S. Dist. LEXIS 51000, at *25–26 (N.D. Cal. Mar. 26, 2018); Nyabwa v. FaceBook, No. 2:17-CV-24, 2018 U.S. Dist.

LEXIS 13981, at *2 (S.D. Tex. Jan. 26, 2018); Quigley v. Yelp, Inc., No. 17-cv-03771-RS, 2017 U.S. Dist. LEXIS

103771, at *4 (N.D. Cal. July 5, 2017); Buza v. Yahoo!, Inc., No. C 11-4422 RS, 2011 U.S. Dist. LEXIS 122806, at *2

(N.D. Cal. Oct. 24, 2011); Young v. Facebook, Inc., No. 5:10-cv-03579-JF/PVT, 2010 U.S. Dist. LEXIS 116530, at *9

(N.D. Cal. Oct. 25, 2010); Estavillo v. Sony Comput. Entm’t Am. Inc., No. C-09-03007 RMW, 2009 U.S. Dist. LEXIS

86821, at *3–4 (N.D. Cal. Sept. 22, 2009); Langdon v. Google, Inc., 474 F. Supp. 2d 622, 631–32 (D. Del. 2007); Nat’l

A-1 Advert. v. Network Solutions, Inc., 121 F. Supp. 2d 156, 169 (D.N.H. 2000); Cyber Promotions v. Am. Online,

948 F. Supp. 436, 445 (E.D. Penn. 1996).

64 See, e.g., Green, 318 F.3d at 472; Nyabwa, 2018 U.S. Dist. LEXIS 13981, at *2; Forbes v. Facebook, Inc., No. 16CV-404 (AMD), 2016 U.S. Dist. LEXIS 19857, at *4–5 (E.D.N.Y. Feb. 18, 2016); Buza, 2011 U.S. Dist. LEXIS

122806, at *2; Langdon, 474 F. Supp. at 631–32.

65 Cyber Promotions, 948 F. Supp. at 441; accord Quigley, 2017 U.S. Dist. LEXIS 103771, at *4.

66 See Prager Univ., 2018 U.S. Dist. LEXIS 51000, at *18; Estavillo, 2009 U.S. Dist. LEXIS 86821, at *4.

67 Prager Univ., 2018 U.S. Dist. LEXIS 51000, at *24; accord Estavillo, 2009 U.S. Dist. LEXIS 86821, at *5.

Congressional Research Service

7

Free Speech and the Regulation of Social Media Content

subject to the First Amendment.68 Courts have rejected plaintiffs’ efforts to characterize the

provision of a public forum69 or “the dissemination of news and fostering of debate”70 as public

functions that were traditionally and exclusively performed by the government.

For example, in Cyber Promotions v. American Online (AOL), a district court rejected the

argument that “by providing Internet e-mail and acting as the sole conduit to its members’

Internet e-mail boxes, AOL has opened up that part of its network [to the public] and as such, has

sufficiently devoted this domain for public use.”71 The court said that “[a]lthough AOL has

technically opened its e-mail system to the public by connecting with the Internet, AOL has not

opened its property to the public by performing any municipal power or essential public service

and, therefore, does not stand in the shoes of the State.”72 The challengers in that case, a company

that had been blocked from sending unsolicited advertisements via email, also argued that AOL

performed an exclusive public function because the company had “no alternative avenues of

communication . . . to send its e-mail to AOL members.”73 The judge rejected this claim as well,

concluding that the company did have alternative avenues to send its advertising to AOL

members, including other places on the internet as well as “non-Internet avenues.”74 Similarly, in

Prager University v. Google LLC, a district court held that by operating YouTube, “a ‘videosharing website,’” and then restricting access to some videos, Google had not “somehow engaged

in one of the ‘very few’ functions that were traditionally ‘exclusively reserved to the State.’”75

Trial courts have also held that social networks have failed to meet the joint participation, nexus,

and entwinement tests for state action.76 In Cyber Promotions, the court held that there was no

joint participation because the government was not involved in AOL’s challenged decision.77

Another trial court, in Quigley v. Yelp, Inc., similarly concluded joint participation did not exist

68 See Prager Univ., 2018 U.S. Dist. LEXIS 51000, at *23–24; Quigley, 2017 U.S. Dist. LEXIS 103771, at *4;

Estavillo, 2009 U.S. Dist. LEXIS 86821, at *4; Cyber Promotions, 948 F. Supp. at 442.

69 E.g., Prager Univ., 2018 U.S. Dist. LEXIS 51000, at *26 (“Defendants do not appear to be at all like, for example, a

private corporation . . . that has been given control over a previously public sidewalk or park . . . .”).

70 Quigley, 2017 U.S. Dist. LEXIS 103771, at *4 (“The dissemination of news and fostering of debate cannot be said to

have been traditionally the exclusive prerogative of the government.”).

71 948 F. Supp. at 442 (internal quotation mark omitted).

72 Id. (emphasis added).

73 Id. at 442–43.

74 Id. at 443.

75 Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 U.S. Dist. LEXIS 51000, at *16 (N.D. Cal. Mar. 26,

2018) (citation omitted) (quoting Flagg Bros., Inc. v. Brooks, 436 U.S. 149, 158 (1978)).

76 Quigley, 2017 U.S. Dist. LEXIS 103771, at *5–7; Estavillo v. Sony Comput. Entm’t Am. Inc., No. C-09-03007

RMW, 2009 U.S. Dist. LEXIS 86821, at *5 (N.D. Cal. Sept. 22, 2009); Cyber Promotions, 948 F. Supp. at 444–45. See

also Fehrenbach v. Zeldin, No. 17-CV-5282, 2018 U.S. Dist. LEXIS 132992, at *7–8 (E.D.N.Y. Aug. 6, 2018) (“[N]o

reasonable person could infer from the pleading that the state ‘encouraged’ or ‘coerced’ the Facebook defendants to

enable users to delete posts or to ‘look away’ if and when they do.”). Cf. Daphne Keller, Who Do You Sue?, HOOVER

INST., Aegis Series Paper No. 1902, at 3 (2019) (“Governments influence platforms’ content removal decisions in a

number of ways.”). However, lower courts have sometimes concluded that the social media profiles of government

agencies or government officials may be public fora in which the First Amendment applies, if they have been opened to

the public. See, e .g., Davison v. Randall, 912 F.3d 666, 682–84 (4th Cir. 2019); Knight First Amendment Inst. v.

Trump, 302 F. Supp. 3d 541, (S.D.N.Y. 2018), appeal filed, No. 18-1691 (2d Cir. June 5, 2018); but see, e.g., Hargis v.

Bevin, No. 3:17-cv-00060-GFVT, 2018 U.S. Dist. LEXIS 54428, at *15 (E.D. Ky. Mar. 30, 2018) (holding that the

First Amendment’s forum doctrine does not apply because state governor’s “use of privately owned Facebook Page and

Twitter pages is personal speech,” and he was “speaking on his own behalf”).

77 948 F. Supp. at 444. The court said that the court’s involvement alone did not implicate the First Amendment. Id. at

444–45.

Congressional Research Service

8

Free Speech and the Regulation of Social Media Content

between various social media sites and the government where the plaintiff failed to show that the

state participated in the specific actions challenged in the lawsuit.78 That court also rejected an

argument that there was “a pervasive entwinement between defendants and the government

because the government maintains accounts on the defendants’ websites, and uses their websites

to communicate with citizens.”79 Even assuming that this allegation was true, the court held that

this was not “the sort of entwinement that . . . converts a private party’s actions to state action,”80

observing that the government did not participate “in the operation or management of defendants’

websites,” but only used these sites “in the same manner as other users.”81

Accordingly, lower courts have uniformly concluded that the First Amendment does not prevent

social media providers from restricting users’ ability to post content on their networks. However,

the Supreme Court has not yet weighed in on this subject, and as will be discussed in more detail

below,82 a number of legal commentators have argued that, notwithstanding these trial court

decisions, courts should view social media platforms as equivalent to state actors, at least when

they perform certain functions.

Section 230 of the CDA

A constitutional injury is not the only type of harm that a social media user might suffer as a

result of a social network’s decisions about user content, and litigants have brought a wide variety

of claims challenging these sorts of decisions. For example, plaintiffs have argued that sites’

decisions to remove or restrict access to their content constituted unfair competition under the

Lanham Act,83 discrimination under the Civil Rights Act of 1964,84 tortious interference with

contractual relationships,85 fraud,86 and breach of contract.87 Other plaintiffs have attempted to

hold online platforms liable for harm stemming from the sites’ decisions not to remove content,

78 Quigley, 2017 U.S. Dist. LEXIS 103771, at *5–6.

79 Id. at *6.

80 Id.

81 Id. at *7. See also Estavillo, 2009 U.S. Dist. LEXIS 86821, at *5 (noting that Sony’s Playstation 3 network “was not

created to further government objectives,” that the government had no authority in the governing of the network, and

that the government did not “encourag[e] Sony to create the Network”).

82 See infra “Social Media Sites as Company Towns.”

83 e-ventures Worldwide, LLC v. Google, Inc., No. 2:14-cv-646-FtM-PAM-CM, 2017 U.S. Dist. LEXIS 88650, at *4

(M.D. Fla. Feb. 8, 2017); Spy Phone Labs LLC v. Google Inc., No. 15-cv-03756-KAW, 2016 U.S. Dist. LEXIS

143530, at *8 (N.D. Cal. Oct. 14, 2016).

84 Sikhs for Justice “SJF”, Inc., v. Facebook, Inc., 144 F. Supp. 3d 1088, 1090 (N.D. Cal. 2015), aff’d, 697 Fed. Appx.

526, 526 (9th Cir. 2017).

85 Zango, Inc. v. Kaspersky Lab, Inc., 568 F.3d 1169, 1172 (9th Cir. 2009); e-ventures Worldwide, LLC, 2017 U.S.

Dist. LEXIS 88650, at *4; Spy Phone Labs LLC, 2016 U.S. Dist. LEXIS 143530, at *8; Search King, Inc. v. Google

Tech., Inc., No. CIV-02-1457-M, 2003 U.S. Dist. LEXIS 27193, at *4 (W.D. Okla. May 27, 2003); cf. e360Insight,

LLC v. Comcast Corp., 546 F. Supp. 2d 605, 607 (N.D. Ill. 2008) (tortious interference with prospective economic

advantage).

86 Langdon v. Google, Inc., 474 F. Supp. 2d 622, 626 (D. Del. 2007). Plaintiffs suing social media companies have also

alleged state law causes of action for deceptive business and trade practices, e-ventures Worldwide, LLC, 2017 U.S.

Dist. LEXIS 88650, at *4; e360Insight, LLC, 546 F. Supp. 2d at 607; Langdon, 474 F. Supp. 2d at 626, and breaches of

covenants of good faith and fair dealing, Darnaa, LLC v. Google, Inc., No. 15-cv-03221-RMW, 2016 U.S. Dist. LEXIS

152126, at *8 (N.D. Cal. Nov. 2, 2016); Spy Phone Labs LLC, 2016 U.S. Dist. LEXIS 143530, at *8; Sikhs for Justice,

144 F. Supp. 3d at 1091.

87 Sikhs for Justice, 144 F. Supp. 3d at 1091; Smith v. Trusted Universal Standards in Elec. Transactions, Inc., No. 094567, 2011 U.S. Dist. LEXIS 26757, at *6 (D.N.J. Mar. 15 2011); Langdon, 474 F. Supp. 2d at 626.

Congressional Research Service

9

Free Speech and the Regulation of Social Media Content

claiming, for example, that by publishing certain content, the sites committed defamation88 or

negligence,89 or violated state securities law.90 However, many of these suits are barred by the

broad grant of immunity created by the CDA’s Section 230.91

47 U.S.C. § 230(c)

(1) TREATMENT OF PUBLISHER OR SPEAKER.—No provider or user of an interactive computer

service shall be treated as the publisher or speaker of any information provided by another information

content provider.

(2) CIVIL LIABILITY.—No provider or user of an interactive computer service shall be held liable on

account of—

(A) any action voluntarily taken in good faith to restrict access to or availability of

material that the provider or user considers to be obscene, lewd, lascivious, filthy,

excessively violent, harassing, or otherwise objectionable, whether or not such

material is constitutionally protected; or

(B) any action taken to enable or make available to information content providers or

others the technical means to restrict access to material described in paragraph (1).92

Section 230, as seen in the text box above, distinguishes between “interactive computer services”

and “information content providers.”93 An interactive computer service is “any information

service, system, or access software provider that provides or enables computer access by multiple

users to a computer server.”94 Courts have considered online platforms such as Facebook,95

Twitter,96 and Craigslist97 to be “interactive computer service” providers.98 An information

content provider is “any person or entity that is responsible, in whole or in part, for the creation or

Barnes v. Yahoo!, Inc., 570 F.3d 1096, 1101 (9th Cir. 2009) (“The cause of action most frequently associated with

. . . section 230 is defamation.”); see, e.g., Ricci v. Teamsters Union Local 456, 781 F.3d 25, 27 (2d Cir. 2015).

89 E.g., Klayman v. Zuckerberg, 753 F.3d 1354, 1359 (D.C. Cir. 2014) (negligence and assault); Barnes, 570 F.3d at

1102-03 (negligent undertaking).

90 Universal Commc’n Sys., Inc. v. Lycos, Inc., 478 F.3d 413, 422 (1st Cir. 2007).

91 47 U.S.C. § 230. See also Universal Commc’n Sys., Inc., 478 F.3d at 418 (“The other courts that have addressed

these issues have generally interpreted Section 230 immunity broadly . . .”); id. at 419 (“[W]e too find that Section 230

immunity should be broadly construed.”). For example, relying on Section 230, courts have dismissed claims alleging

that various social media platforms have supported terrorist organizations by allowing terrorists to use their sites. See

Cohen v. Facebook, Inc., 252 F. Supp. 3d 140, 145 (E.D.N.Y. 2017); Fields v. Twitter, Inc., 217 F. Supp. 3d 1116,

1118 (N.D. Cal. 2016), aff’d on other grounds, 881 F.3d 739, 750 (9th Cir. 2018).

92 Courts have read 47 U.S.C. § 230(c)(2)(B)’s reference to “paragraph (1)” to mean § 230(c)(2)(A). E.g. Zango, Inc. v.

Kaspersky Lab, Inc., 568 F.3d 1169, 1173 n.5 (9th Cir. 2009) (“We take it that the reference to the ‘material described

in paragraph (1)’ is a typographical error, and that instead the reference should be to paragraph (A), i.e., § 230(c)(2)(A).

Paragraph (1) pertains to the treatment of a publisher or speaker and has nothing to do with ‘material,’ whereas

subparagraph (A) pertains to and describes material.”) (citation omitted).

93 47 U.S.C. § 230(f).

94 Id. § 230(f)(2).

95 E.g., Klayman v. Zuckerberg, 753 F.3d 1354, 1357 (D.C. Cir. 2014).

96 E.g., Fields v. Twitter, 217 F. Supp. 3d 1116, 1121 (N.D. Cal. 2016).

97 E.g., Chicago Lawyers’ Comm. for Civil Rights under Law, Inc. v. Craigslist, Inc., 519 F.3d 666, 671 (7th Cir.

2008).

98 See also Universal Commc’n Sys., Inc. v. Lycos, Inc., 478 F.3d 413, 419 (1st Cir. 2007) (“Providing access to the

Internet is . . . not the only way to be an interactive computer service provider.”).

88

Congressional Research Service

10

Free Speech and the Regulation of Social Media Content

development of information provided through the Internet or any other interactive computer

service.”99

Section 230 contains two primary provisions creating immunity from liability. First, Section

230(c)(1) specifies that interactive service providers and users may not “be treated as the

publisher or speaker of any information provided by another information content provider.”100

Second, Section 230(c)(2) states that interactive service providers and users may not be held

liable for voluntarily acting in good faith to restrict access to objectionable material.101 Section

230 preempts state civil lawsuits and state criminal prosecutions to the extent that they are

“inconsistent” with Section 230.102 It also bars certain federal civil lawsuits,103 but, significantly,

not federal criminal prosecutions.104 Section 230(e) outlines a few exemptions: for example,

Section 230 immunity will not apply in a suit “pertaining to intellectual property”105 or in claims

alleging violations of certain sex trafficking laws.106

Section 230(c)(1)

Section 230, and particularly Section 230(c)(1), distinguishes those who create content from those

who provide access to that content, providing immunity to the latter group.107 An entity may be

both an “interactive computer service” provider and an “information content provider,” but the

critical inquiry for applying Section 230(c)(1) is whether, with respect to the particular actions

alleged to create liability, the service provider developed the underlying content.108

Courts have held that an interactive computer service provider may be subject to suit if it is also

acting as a content provider.109 Frequently, the application of Section 230(c)(1) immunity turns

99 47 U.S.C. § 230(f)(3).

100 Id. § 230(c)(1).

101 Id. § 230(c)(2).

47 U.S.C. § 230(e)(3) preempts contrary state law, providing that “[n]o cause of action may be brought and no

liability may be imposed under any State or local law that is inconsistent with this section.” See also, e.g., Doe v. GTE

Corp., 347 F.3d 655, 658 (7th Cir. 2003) (“These provisions preempt contrary state law.”); Backpage.com, LLC v.

Cooper, 939 F. Supp. 2d 805, 823 (M.D. Tenn. 2013) (concluding that a state law imposing criminal liability “is likely

expressly preempted by” the CDA).

103 See, e.g., Fair Hous. Council v. Roommates.com, LLC, 521 F.3d 1157, 1174–75 (9th Cir. 2008).

104 47 U.S.C. § 230(e)(1) provides that Section 230 shall not “be construed to impair the enforcement of . . . any . . .

Federal criminal statute.”

105 Id. § 230(e)(2) (“Nothing in this section shall be construed to limit or expand any law pertaining to intellectual

property.”); see, e.g., Gucci Am., Inc. v. Hall & Assocs., 135 F. Supp. 2d 409, 413 (S.D.N.Y. 2001) (holding that

providing immunity from a suit alleging a violation of intellectual property laws “would ‘limit’ the laws pertaining to

intellectual property in contravention of § 230(e)(2)”).

106 See 47 U.S.C. § 230(e)(5).

107 See id. § 230(f). See also, e.g., Ricci v. Teamsters Union Local 456, 781 F.3d 25, 28 (2d Cir. 2015) (“In short, a

plaintiff defamed on the internet can sue the original speaker, but typically ‘cannot sue the messenger.’” (quoting

Chicago Lawyers’ Comm. for Civil Rights under Law, Inc. v. Craigslist, Inc., 519 F.3d 666, 672 (7th Cir. 2008))).

Some have noted, however, that 47 U.S.C. § 230(c)(1) does not itself use the word “immunity,” or any comparable

legal terms. See Chicago Lawyers’ Comm. for Civil Rights under Law, Inc., 519 F.3d at 669; Doe v. GTE Corp., 347

F.3d 655, 660–61 (7th Cir. 2003). Compare Barnes v. Yahoo!, Inc., 570 F.3d 1096, 1100–01 (9th Cir. 2009)

(explaining that 47 U.S.C. § 230(c)(1) alone does not create immunity from suit, but creates protection from liability

when viewed in conjunction with 47 U.S.C. § 230(e)(3)), with Zeran v. Am. Online, Inc., 129 F.3d 327, 330 (4th Cir.

1997) (“By its plain language, [47 U.S.C. § 230(c)(1)] creates a federal immunity to any cause of action that would

make service providers liable for information originating with a third-party user of the service.”).

108 See, e.g., Fair Hous. Council v. Roommates.com, LLC, 521 F.3d 1157, 1174 (9th Cir. 2008).

109 Id. at 1162-63. Some circuits have described a three-part test to determine whether an entity posting information is

102

Congressional Research Service

11

Free Speech and the Regulation of Social Media Content

not on the type of suit that is being brought—that is, for example, whether it is a suit for libel or

for breach of contract110—but on whether the facts establish that the interactive computer service

provider was merely a publisher of another’s content,111 or whether the service provider itself

created or developed content.112 Courts have generally held that a site’s ability to control the

content posted on its website does not, in and of itself, transform an interactive computer service

into an internet content provider.113 As one court said, “a website does not create or develop

content when it merely provides a neutral means by which third parties can post information of

their own independent choosing online.”114 A service provider may still be immune from suit

under Section 230(c)(1) even if it makes small editorial changes to that content.115

Conversely, a “website operator” can be liable for “content that it creates itself, or is ‘responsible,

in whole or in part’ for creating or developing.”116 Even if the service provider does not itself

solely create the content, Section 230 immunity might be unavailable if the service provider

“augment[s] the content.”117 For example, one state court held that, even assuming that Snapchat

was a provider of interactive computer services, a plaintiff’s claim against the company could

proceed where the alleged harm was caused by a “filter,” or a graphic overlay on a user’s photo,

that was created by Snapchat itself.118 Because the plaintiff sought “to hold Snapchat liable for its

own conduct,” the court held that “CDA immunity does not apply.”119

Some courts have applied a “material contribution test,”120 asking whether a service provider

“materially contribute[d] to the illegality” of the disputed content,121 or “in some way specifically

encourage[d] development of what is offensive about the content.”122 Thus, for example, a federal

entitled to immunity: Section 230 “protects from liability (1) a provider or user of an interactive computer service

(2) whom a plaintiff seeks to treat . . . as a publisher or speaker (3) of information provided by another information

content provider.” Barnes, 570 F.3d at 1100–01. Accord Jones v. Dirty World Entm’t Recordings, LLC, 755 F.3d 398,

409 (6th Cir. 2014); Klayman v. Zuckerberg, 753 F.3d 1354, 1357 (D.C. Cir. 2014); Fed. Trade Comm’n v.

Accusearch, Inc., 570 F.3d 1187, 1196 (10th Cir. 2009).

110 See, e.g., Barnes, 570 F.3d at 1102–03 (“[W]hat matters is not the name of the cause of action . . . [but] whether the

cause of action inherently requires the court to treat the defendant as the ‘publisher or speaker’ of content provided by

another.”). Cf., e.g., Doe v. Internet Brands, Inc., 824 F.3d 846, 851, 853 (9th Cir. 2016) (concluding Section 230 did

not bar plaintiff’s claim because the type of tort duty allegedly breached was not within the scope of the statute).

111 See 47 U.S.C. § 230(f)(2).

112 See Fed. Trade Comm’n v. LeadClick Media, LLC, 838 F.3d 158, 174, 176 (2d Cir. 2016). See also Universal

Commc’n Sys., Inc. v. Lycos, Inc., 478 F.3d 413, 418 (1st Cir. 2007) (“[A]n interactive computer service provider

remains liable for its own speech.”).

113 See, e.g., Klayman, 753 F.3d at 1358.

114 Id.

115 See Batzel v. Smith, 333 F.3d 1018, 1031 (9th Cir. 2003); Ben Ezra, Weinstein, & Co. v. Am. Online, Inc., 206 F.3d

980, 985 (10th Cir. 2000). See also Doe v. Friendfinder Network, Inc., 540 F. Supp. 2d 288, 299 (D.N.H. 2008) (“[A]

service provider’s privilege as a ‘publisher’ under the Act protects more than the mere repetition of data obtained from

another source, but extends to the provider’s ‘inherent decisions about how to treat postings generally.’”) (quoting

Universal Commc’n Sys., Inc., 478 F.3d at 422).

116 Fair Hous. Council v. Roommates.com, LLC, 521 F.3d 1157, 1170, 1162 (9th Cir. 2008) (quoting 47 U.S.C.

§ 230(f)(3)).

117 Id. at 1167–68.

118 Maynard v. Snapchat, Inc., 816 S.E.2d 77, 79–81 (Ga. Ct. App. 2018).

119 Id. at 81.

120 Jones v. Dirty World Entm’t Recordings, LLC, 755 F.3d 398, 413 (6th Cir. 2014).

121 Id. at 415. Accord, e.g., Fair Hous. Council, 521 F.3d at 1168.

122 Fed. Trade Comm’n v. Accusearch, Inc., 570 F.3d 1187, 1199 (10th Cir. 2009). Cf. Jones, 755 F.3d at 414–15

(rejecting “an encouragement theory of ‘development’” that would impose liability for intentionally encouraging

Congressional Research Service

12

Free Speech and the Regulation of Social Media Content

appellate court concluded that Roommates.com, a site that “match[ed] people renting out spare

rooms with people looking for a place to live,” was not wholly immune from claims that it had

violated laws prohibiting housing discrimination.123 The court concluded that Roommates.com

could be subject to suit for discrimination because the site required all users to respond to

questions about their sex, family status, and sexual orientation by selecting among preset answers

to those questions, and to state their “preferences in roommates with respect to the same three

criteria.”124 Accordingly, in the court’s view, as to these questions and answers, Roommates.com

was “more than a passive transmitter of information provided by others; it becomes the developer,

at least in part, of that information.”125 Each user’s personal page was “a collaborative effort

between [Roommates.com] and the subscriber.”126 This rendered it the “‘information content

provider’ as to the questions”127 and the answers.128

Section 230(c)(2)

Although courts frequently consider the immunity in Section 230(c)(1) and Section 230(c)(2)

together, as one “Section 230” shield,129 the text of these provisions suggests they cover distinct

circumstances. Section 230(c)(1) applies more broadly, to any suit in which the plaintiff seeks to

hold the provider liable as the publisher of another’s information.130 By contrast, Section

230(c)(2) applies only to good-faith, voluntary actions by a provider—or a third party assisting

providers—to restrict access to “obscene, lewd, lascivious, filthy, excessively violent, harassing,

or otherwise objectionable” content.131

There is an important difference in the language of the two provisions: as noted, Section

230(c)(2) requires a service provider to act in good faith for immunity to apply; Section 230(c)(1)

does not contain a similar requirement.132 While courts frequently apply Section 230 to dismiss

lawsuits premised on a service provider’s decision to remove or restrict access to another’s

content,133 they are somewhat less likely to dismiss lawsuits where the good-faith requirement is

involved, because a plaintiff who properly pleads and presents evidence regarding a lack of good

certain types of third-party postings).

123 Fair Hous. Council, 521 F.3d at 1161, 1175.

124 Id. at 1161.

125 Id.

126 Id. at 1167.

127 Id. at 1164 (quoting 47 U.S.C. § 230(f)(3)).

128 Id. at 1166.

129 See, e.g., Zeran v. Am. Online, Inc., 129 F.3d 327, 330 (4th Cir. 1997). See also Eric Goldman, MySpace Quietly

Won Goofy 230 Ruling in September–Riggs v. MySpace, TECH. & MKTG. LAW BLOG (Nov. 30, 2009),

https://bit.ly/2S3CMrW (noting that courts have “collapsed the two provisions together . . . with surprising regularity”)

[hereinafter Goldman, Riggs v. MySpace].

130 47 U.S.C. § 230(c)(1); Klayman v. Zuckerberg, 753 F.3d 1354, 1357 (D.C. Cir. 2014).

131 47 U.S.C. § 230(c)(2).

132 See id. § 230(c).

133 Zango, Inc. v. Kaspersky Lab, Inc., 568 F.3d 1169, 1174 (9th Cir. 2009); Mezey v. Twitter, Inc., No. 1:18-cv21069-KMM, 2018 U.S. Dist. LEXIS 121775, at *3 (S.D. Fla. July 19, 2018); Darnaa, LLC v. Google, Inc., No. 15-cv03221-RMW, 2016 U.S. Dist. LEXIS 152126, at *23 (N.D. Cal. Nov. 2, 2016); Sikhs for Justice “SJF”, Inc., v.

Facebook, Inc., 144 F. Supp. 3d 1088, 1095 (N.D. Cal. 2015), aff’d, 697 Fed. Appx. 526, 526 (9th Cir. 2017); Smith v.

Trusted Universal Standards in Elec. Transactions, Inc., No. 09-4567, 2011 U.S. Dist. LEXIS 26757, at *16–24 (D.N.J.

Mar. 15 2011); Riggs v. MySpace, No. CV 09-03073-GHK (CTx), 2009 WL 10671689, at *3 (C.D. Cal. Sept. 17,

2009), aff’d, 444 Fed. Appx. 986, 987 (9th Cir. 2011); e360Insight, LLC v. Comcast Corp., 546 F. Supp. 2d 605, 609

(N.D. Ill. 2008); Langdon v. Google, Inc., 474 F. Supp. 2d 622, 631 (D. Del. 2007).

Congressional Research Service

13

Free Speech and the Regulation of Social Media Content

faith creates a question of fact that may prevent the court from summarily dismissing the case.134

One trial court concluded that there was a question as to Google’s good faith where the plaintiff

alleged that Google was “selectively enforcing” a stated policy, and that the policy itself was

“entirely pretextual.”135 Another trial court concluded that a company had sufficiently alleged bad

faith where it argued that Google had “falsely accused” it of violating Google’s stated policy, and

that Google “sought to punish [the company] because it” refused to allow Google to embed

advertising in the company’s video.136

One view is that Section 230(c)(2) applies when a provider “does filter out offensive material,”

while Section 230(c)(1) applies when providers “refrain from filtering or censoring the

information on their sites.”137 At least one federal trial judge has noted that interpreting Section

230(c)(1) to bar suits in which a plaintiff seeks to hold a service provider liable for removing the

plaintiff’s own content would “swallow[] the more specific immunity in (c)(2).”138 The court

explained that:

Subsection (c)(2) immunizes only an interactive computer service’s “actions taken in good

faith.” If the publisher’s motives are irrelevant and always immunized by (c)(1), then (c)(2)

is unnecessary. The Court is unwilling to read the statute in a way that renders the goodfaith requirement superfluous. 139

Lawsuits directly challenging a website’s decision to restrict or remove content, rather than to

publish it, often do invoke Section 230(c)(2).140 Thus, courts have considered the application of

Section 230(c)(2), rather than Section 230(c)(1), in lawsuits involving the removal of an app from

the Google Play Store,141 the removal of websites from Google’s search results,142 the removal of

videos from YouTube,143 and decisions to filter certain IP addresses or email addresses.144

However, this distinction between filtering content and publishing content does not always play

134

See, e.g., e-ventures Worldwide, LLC v. Google, Inc., No. 2:14-cv-646-FtM-PAM-CM, 2017 U.S. Dist. LEXIS

88650, at *4 (M.D. Fla. Feb. 8, 2017); Darnaa, LLC, 2016 U.S. Dist. LEXIS 152126, at *25; Spy Phone Labs LLC v.

Google Inc., No. 15-cv-03756-KAW, 2016 U.S. Dist. LEXIS 143530, at *26 (N.D. Cal. Oct. 14, 2016); Smith, 2011

U.S. Dist. LEXIS 26757, at *25–26. But see, e.g., Zango, Inc., 568 F.3d at 1174 (applying 47 U.S.C. § 230(c)(2) to

dismiss claim); Smith, 2011 U.S. Dist. LEXIS 26757, at *16–24 (same); e360Insight, LLC, 546 F. Supp. 2d at 609

(same); Langdon, 474 F. Supp. 2d at 631 (same).

135 Spy Phone Labs LLC, 2016 U.S. Dist. LEXIS 143530, at *26.

136 Darnaa, LLC, 2016 U.S. Dist. LEXIS 152126, at *25–26.

137 Doe v. GTE Corp., 347 F.3d 655, 659 (7th Cir. 2003).

138 e-ventures Worldwide, LLC, 2017 U.S. Dist. LEXIS 88650, at *9.

139 Id.

140 See, e.g., Smith v. Trusted Universal Standards in Elec. Transactions, Inc., No. 09-4567, 2011 U.S. Dist. LEXIS

26757, at *16–24 (D.N.J. Mar. 15 2011); e360Insight, LLC v. Comcast Corp., 546 F. Supp. 2d 605, 607 (N.D. Ill.

2008); Langdon v. Google, Inc., 474 F. Supp. 2d 622, 630 (D. Del. 2007). Compare Green v. Am. Online (AOL), 318

F.3d 465, 470 (3rd Cir. 2003) (holding that Section 230(c)(1) immunity applies to AOL’s “alleged negligent failure to

properly police its network for content transmitted by its users”), with id. at 473 (holding that Section 230(c)(2)

immunity applies to AOL’s allegedly unlawful conduct preventing users from sending objectionable emails).

141 Spy Phone Labs LLC v. Google Inc., No. 15-cv-03756-KAW, 2016 U.S. Dist. LEXIS 143530, at *25–26 (N.D. Cal.

Oct. 14, 2016).

142 e-ventures Worldwide, LLC, 2017 U.S. Dist. LEXIS 88650, at *4.

143 Darnaa, LLC v. Google, Inc., No. 15-cv-03221-RMW, 2016 U.S. Dist. LEXIS 152126, at *25 (N.D. Cal. Nov. 2,

2016).

144 Zango, Inc. v. Kaspersky Lab, Inc., 568 F.3d 1169, 1174 (9th Cir. 2009); Smith, 2011 U.S. Dist. LEXIS 26757, at

*16–26.

Congressional Research Service

14

Free Speech and the Regulation of Social Media Content

out so neatly in the courts, and other decisions have applied Section 230(c)(1) immunity to bar

suits that are grounded in an interactive service provider’s decision to restrict content.145

There is one additional circumstance under which Section 230(c)(2) immunity, as opposed to

Section 230(c)(1) immunity, may apply. Section 230(c)(2)(B) protects those providers or users of

computer services who “enable or make available to information content providers or others

technical means to restrict access to” objectionable material.146 This provision may protect, for

example, “providers of programs that filter adware and malware.”147 This immunity may apply

even where an interactive computer service is not a publisher entitled to immunity under Section

230(c)(1).148

Thus, as a whole, Section 230 offers broad immunity to “interactive computer service” providers

when a litigant seeks to hold them liable for publishing, or not publishing, a user’s content.149

Section 230(c)(1) provides immunity from any lawsuit that seeks to hold a service provider liable

for publishing information that was created by an “information content provider,” effectively

protecting social media sites from liability for hosting content.150 And Section 230(c)(2) provides

immunity for sites that take good faith action to restrict access to content that the provider or

users deem “objectionable.”151 Consequently, to the extent that private litigants or state

governments would have been able to hold social media companies liable under existing law for

their decisions regarding presenting or restricting access to user content, those suits have largely

been barred under Section 230.152

First Amendment Limits on Government Regulation

of Social Media Content

As discussed above, courts have often dismissed lawsuits attempting to hold social media

providers liable for regulating users’ content, whether because the court concludes that the First

Amendment does not apply to the actions of these private actors or because the court holds that

Section 230(c)(2) of the CDA bars the lawsuit. Additionally, Section 230(c)(1) may bar lawsuits

145 See, e.g., Mezey v. Twitter, Inc., No. 1:18-cv-21069-KMM, 2018 U.S. Dist. LEXIS 121775, at *2–3 (S.D. Fla. July

19, 2018); Sikhs for Justice Inc. v. Facebook, Inc., 144 F. Supp. 3d 1088, 1096 (N.D. Cal. 2015), aff’d, 697 Fed. Appx.

526, 526 (9th Cir. 2017); Riggs v. MySpace, No. CV 09-03073-GHK (CTx), 2009 WL 10671689, at *3 (C.D. Cal.

Sept. 17, 2009), aff’d, 444 Fed. Appx. 986, 987 (9th Cir. 2011). And perhaps somewhat confusingly, some cases

applying Section 230(c)(2) immunity cite, as part of their justification, cases that describe publisher immunity under

Section 230(c)(1). See, e.g., Langdon v. Google, Inc., 474 F. Supp. 2d 622, 631 (D. Del. 2007).

146 47 U.S.C. § 230(c)(2)(B).

147 Zango, Inc. v. Kaspersky Lab, Inc., 568 F.3d 1169, 1174 (9th Cir. 2009). See generally, e.g., Russell A. Miller, The

Legal Fate of Internet Ad-Blocking, 24 B.U. J. SCI. & TECH. L. 301, 358–60 (2018). Section 230(c)(2)(B) has also been

applied to entities like Facebook. Fehrenbach v. Zeldin, No. 17-CV-5282, 2018 U.S. Dist. LEXIS 132992, at *14

(E.D.N.Y. Aug. 6, 2018). By contrast, one court concluded that Yahoo! was not entitled to immunity under Section

230(c)(2)(B) where it “did not engage in any form of content analysis of the subject text to identify material that was

offensive or harmful.” Sherman v. Yahoo! Inc., 997 F. Supp. 2d 1129, 1138 (S.D. Cal. 2014).

148 See, e.g., Smith v. Trusted Universal Standards in Elec. Transactions, Inc., No. 09-4567, 2011 U.S. Dist. LEXIS

26757, at *15 (D.N.J. Mar. 15 2011).

149 47 U.S.C. § 230(c).

150 Id. § 230(c)(1).

151 Id. § 230(c)(2).

152 See, e.g., Klayman v. Zuckerberg, 753 F.3d 1354, 1358–59 (D.C. Cir. 2014) (discussing scope of immunity

provided by Section 230).

Congressional Research Service

15

Free Speech and the Regulation of Social Media Content

that seek to hold these platforms liable because of their decisions to publish certain content.

Particularly because of Section 230, there are few, if any, federal or state laws that expressly

govern social media sites’ decisions about whether and how to present users’ content.

Consequently, users’ ability to post speech on social media platforms is governed primarily by the

private moderation policies created by these companies.153

In response to broader public policy concerns about how social media entities are policing user

content, some commentators and legislators have proposed federal regulation both to protect

users’ ability to speak freely on those platforms154 and to require these platforms to take down,

deemphasize, or clarify certain content.155 While the First Amendment, as discussed above, may

not apply in disputes between private parties,156 a federal law regulating internet content decisions

would likely qualify as state action sufficient to implicate the First Amendment.157 After all, the

First Amendment provides that “Congress shall make no law . . . abridging the freedom of

speech.”158

Once state action is established, the next consideration is to what extent the First Amendment

protects social media platforms’ content moderation decisions. Stated another way, the relevant

question is when social media providers can assert that government regulation infringes on their

own speech. Perhaps most obviously, if a social media site posts content that it has created itself,

the site may raise First Amendment objections to a law expressly regulating that speech.159 Social

media providers may also argue that they are exercising protected speech rights when they are

choosing whether to publish content that was originally created by users and when they make

decisions about how to present that content.160 However, the fact that a law affects speech

153 See, e.g., Klonick, supra note 32, at 1630–58 (describing these content moderation policies).

154 See, e.g., H.R. 492, 116th Cong. (2019) (amending the CDA to provide that “an owner or operator of a social

media service . . . shall be treated as a publisher or speaker of such content” based on certain ways of displaying “usergenerated content”); H.R. 4682, 115th Cong. (2017) (amending the CDA to provide that a person providing “broadband

internet access service . . . may not block lawful content”); David L. Hudson, Jr., In the Age of Social Media, Expand

the Reach of the First Amendment, AM. BAR ASS’N, https://bit.ly/2V7eCP3 (last visited Mar. 27, 2019); see also supra

note 26.

155 See, e.g., H.R. 1, § 4207(b)(1), 116th Cong. (2019) (requiring specified disclosures for certain online political

advertisements); Press Release, Rep. Adam Schiff, Schiff Sends Letter to Google, Facebook Regarding Anti-Vaccine

Misinformation (Feb. 14, 2019), https://bit.ly/2XOQqmX; see also supra note 20.

156 Cf., e.g., Henry H. Perritt, Jr., Tort Liability, The First Amendment, and Equal Access to Electronic Networks, 5

HARV. J.L. & TECH. 65, 128 (1992) (“The First Amendment is not likely to be an effective sword for information

service providers seeking to use it as an affirmative source of equal access rights. Only if the FCC or the Congress or

another governmental entity imposes specific affirmative requirements on private information service providers is the

conduct of the service provider likely to be sufficient state action to implicate the First Amendment.”).

157 See, e.g., Reno v. ACLU, 521 U.S. 844, 882 (1997) (striking down as unconstitutional under the First Amendment

provisions of the CDA that prohibited the transmission of indecent or patently offensive messages to minors); cf., e.g.,

Janus v. Am. Fed’n of State, Cty., and Mun. Emps., Council 31, 138 S. Ct. 2448, 2479 n.24 (2018) (questioning

whether a federal statute “allowing, but not requiring” private third parties to take certain actions is “sufficient to

establish governmental action” under the First Amendment).

158 U.S. CONST. amend. I (emphasis added).

159 Cf., e.g., U.S. Telecom Ass’n v. FCC, 855 F.3d 381, 382 (D.C. Cir. 2017) (Kavanaugh, J., dissenting from denial of

rehearing en banc) (noting that FCC’s net neutrality rule “reflects a fear that the real threat to free speech today comes

from private entities such as Internet service providers,” but rejecting this argument because “the First Amendment is a

restraint on the Government and protects private editors and speakers from Government regulation”); U.S. Telecom

Ass’n v. FCC, 825 F.3d 674, 741–42 (D.C. Cir. 2016) (noting that “insofar as a[n internet] broadband provider might

offer its own content—such as a news or weather site—separate from its internet access service, the provider would

receive the same protection under the First Amendment as other producers of internet content”).

160 E.g., Stuart Minor Benjamin, Determining What “The Freedom of Speech” Encompasses, 60 DUKE L.J. 1673, 1680

(2011). Cf., e.g., Search King, Inc. v. Google Tech., Inc., No. CIV-02-1457-M, 2003 U.S. Dist. LEXIS 27193, at *12

Congressional Research Service

16

Free Speech and the Regulation of Social Media Content

protected by the First Amendment does not necessarily mean that it is unconstitutional. As

explained below, the First Amendment allows some regulation of speech and does not prohibit

regulation of conduct.

Background Principles: First Amendment Protections Online

While the First Amendment generally protects the “freedom of speech,”161 its protections do not

apply in the same way in all cases. Not every government regulation affecting content posted on

social media sites would be analyzed in the same way. A court’s analysis would depend on a

number of factors.

First, a court would inquire into the nature of the precise action being regulated, including

whether it is properly characterized as speech or conduct.162 Laws that target conduct and only

incidentally burden speech may be permissible.163 But “speech” is not always easy to identify.

Lower courts have held that computer code and programs may be entitled to First Amendment

protection, so long as they communicate “information comprehensible to human beings.”164

Courts have also concluded that in some circumstances, domain names might constitute protected

speech.165 And more generally, the Supreme Court has said that “inherently expressive” conduct

can receive First Amendment protections.166

If a law does regulate speech, a court would consider the type of speech being regulated to

determine how closely to scrutinize the regulation. For example, a court may ask whether that

speech is commercial167 and, as such, deserving of less protection under the First Amendment.168

(W.D. Okla. May 27, 2003) (holding that Google’s PageRanks “are constitutionally protected opinions”).

161 U.S. CONST. amend. I.

162 See Ashcroft v. Free Speech Coal., 535 U.S, 234, 253 (2002) (“[T]he Court’s First Amendment cases draw vital

distinctions between words and deeds, between ideas and conduct.”); see also, e.g., Junger v. Daley, 209 F.3d 481,

484–85 (6th Cir. 2000) (noting that encryption source code “has both an expressive feature and a functional feature,”

but ultimately concluding that “[b]ecause computer source code is an expressive means for the exchange of information

and ideas about computer programming, . . . it is protected by the First Amendment”).

163 Rumsfeld v. Forum for Acad. & Institutional Rights, Inc., 547 U.S. 47, 62 (2006) (upholding law where “the

compelled speech . . . is plainly incidental to the [law’s] regulation of conduct”). However, even if a law generally

regulates conduct, it may still be subject to heightened scrutiny if it is applied to speech. See, e.g., Holder v.

Humanitarian Law Project, 561 U.S. 1, 28 (2010).

164 Univ. City Studios, Inc. v. Corley, 273 F.3d 429, 450 (2d Cir. 2001); see also, e.g., Junger v. Daley, 209 F.3d 481,

485 (6th Cir. 2000) (“Because computer source code is an expressive means for the exchange of information and ideas

about computer programming, we hold that it is protected by the First Amendment.”).

165 See, e.g., Name.Space, Inc. v. Network Solutions, Inc., 202 F.3d 573, 586 (2d Cir. 2000) (“[W]hile we hold that the

existing [generic Top Level Domains (gTLDs)] do not constitute protected speech under the First Amendment, we do

not preclude the possibility that certain domain names . . . could indeed amount to protected speech . . . .”); cf. Nat’l A1 Advert. v. Network Sols., Inc., 121 F. Supp. 2d 156, 171 (D.N.H. 2000) (noting that domain names might be

“expressive of some sort of message or idea,” but holding that the challenged names are not protected speech because

“the space occupied by second-level domain names does not constitute a discrete ‘forum’ for speech”).

166 Forum for Acad. & Institutional Rights, Inc., 547 U.S. at 66.

167 The Court has defined purely commercial speech as speech that “does ‘no more than propose a commercial

transaction,’” Va. State Bd. of Pharm. v. Va. Citizens Consumer Council, 425 U.S. 748, 762 (1976) (quoting Pittsburgh

Press Co. v. Human Relations Comm’n, 413 U.S. 376, 385 (1973)), or is “related solely to the economic interests of the

speaker and its audience,” Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n, 447 U.S. 557, 561 (1980).

168 Cent. Hudson, 447 U.S. at 562–63 (“The Constitution . . . accords a lesser protection to commercial speech than to

other constitutionally guaranteed expression.”); see also, e.g., Zhang v. Baidu.com, Inc., 10 F. Supp. 3d 433, 443

(S.D.N.Y. 2014) (holding that decision to block certain search results was not commercial speech, because the “search

results at issue in [that] case . . . relate[d] to matters of public concern and [did] not themselves propose transactions”).

Congressional Research Service

17

Free Speech and the Regulation of Social Media Content

Advertisements posted on social media sites would likely qualify as commercial speech.169 If

speech is not purely commercial and is instead, for example, political advocacy, that speech may

receive greater protection.170 Certain categories of speech receive even less protection than

commercial speech.171 For example, the Supreme Court has said that states may prohibit speech

advocating violence if that “advocacy is directed to inciting or producing imminent lawless action

and is likely to incite or produce such action.”172 Thus, certain types of threatening or violent

speech posted on social media may not be entitled to First Amendment protection. However,

perhaps in light of the fact that it can be difficult to determine whether speech is protected, the

Court has sometimes held that criminal statutes targeting disfavored speech must include a mental

state requirement.173 For example, in United States v. X-Citement Video, the Court noted that, with

respect to a federal law prohibiting the distribution of child pornography,174 criminal liability

turned on “the age of the performers”—as did First Amendment protection for the materials,

given that “nonobscene, sexually explicit materials involving persons over the age of 17 are

protected by the First Amendment.”175 Accordingly, although the statute was unclear on this point,

the Court held that the law applied only if a person distributing such materials knew that the

performers were underage.176

Even if a statute does target a category of speech that is traditionally proscribable, it may still be

invalid if it is overbroad, in the sense that it prohibits a substantial amount of protected speech, as

well.177 Thus, for example, in Ashcroft v. Free Speech Coalition, the Supreme Court held that a

federal statute prohibiting “sexually explicit images that appear to depict minors”178 was

unconstitutionally overbroad.179 The statute encompassed pornography that did “not depict an

actual child,” prohibiting images that were “created by using adults who look like minors or by

169 See Cent. Hudson, 447 U.S. at 563.

170 See, e.g., Citizens United v. FEC, 558 U.S. 310, 340 (2010); Riley v. Nat’l Fed’n of the Blind, 487 U.S. 781, 796

(1988).

171 For example, the First Amendment does not prohibit states from regulating obscenity, defamation, or fighting

words. See, e.g., R.A.V. v. City of St. Paul, 505 U.S. 377, 383 (1992). The Court has frequently described these

categories of speech as unprotected, see, e.g., Miller v. California, 413 U.S. 15, 20–23 (1973), but more recently, the

Court has questioned whether this is technically true, instead emphasizing that these areas of speech may be more

freely regulated, see R.A.V., 505 U.S. at 383–84. Notably, “hate speech” is not, as such, one of these disfavored

categories. See, e.g., Julian Baumrin, Internet Hate Speech & the First Amendment, Revisited, 37 RUTGERS COMPUTER

& TECH. L.J. 223, 226 (2011). See generally CRS In Focus IF11072, The First Amendment: Categories of Speech, by

Victoria L. Killion (describing regulable categories of speech).

172 Brandenburg v. Ohio, 395 U.S. 444, 447 (1969) (per curiam).

173 See, e.g., Elonis v. United States, 135 S. Ct. 2001, 2012 (2015) (holding that a state could not prosecute a person

who posted threats on Facebook where the law did not include a mental state requirement requiring the defendant to

intend to issue a threat or know that the communication would be seen as a threat); United States v. X-Citement Video,

513 U.S. 64, 78 (1994) (concluding that the word “knowingly” in a federal criminal statute extended to both elements

of the crime); Smith v. California, 361 U.S. 147, 155 (1959) (holding that a city could not prohibit bookstores from

possessing obscene books where the law did not require booksellers to know that the books were obscene).

174 As described by the Court, “[t]he Protection of Children Against Sexual Exploitation Act of 1977, as amended,

prohibit[ed] the interstate transportation, shipping, receipt, distribution, or reproduction of visual depictions of minors

engaged in sexually explicit conduct.” 513 U.S. at 65–66.

175 Id. at 72–73.

176 Id. at 78 (“[W]e conclude that the term ‘knowingly’ in [18 U.S.C.] § 2252 extends both to the sexually explicit

nature of the material and to the age of the performers.”).

177 See, e.g., Broadrick v. Oklahoma, 413 U.S. 601, 615 (1973).

178 535 U.S. 234, 239 (2002).

179 Id. at 256.

Congressional Research Service

18

Free Speech and the Regulation of Social Media Content

using computer imaging.”180 Thus, the Court held that the statute violated the First Amendment

because it “proscribe[d] a significant universe of speech that is neither obscene . . . nor child

pornography,” as those two categories had been defined in prior Supreme Court cases.181

A court would also look to the nature of the regulation itself, and primarily whether it is contentneutral, or whether it instead discriminates on the basis of content or viewpoint, subjecting that

law to strict scrutiny.182 The Court has said that “a speech regulation is content based if the law

applies to particular speech because of the topic discussed or the idea or message expressed.”183

In a strict scrutiny analysis, the government must prove that the “restriction ‘furthers a

compelling interest and is narrowly tailored to achieve that interest.’”184 If a regulation is contentneutral, which is to say, “justified without reference to the content of the regulated speech,” a

court employs an intermediate scrutiny analysis, asking whether the restriction is “narrowly

tailored to serve a significant governmental interest” and “leave[s] open ample alternative

channels for communication of the information.”185 Accordingly, for example, a federal court of

appeals held in Universal City Studios, Inc. v. Corley that government restrictions on posting or

linking to decryption computer programs did regulate speech protected by the First Amendment,

but ultimately upheld those restrictions as permissible content-neutral regulations.186

These first two inquiries are distinct, although they do overlap. If a statute targets speech rather

than conduct, it is likely that it will target that speech based on its content, and therefore will not

be content-neutral.187 And by contrast, a statute that targets conduct will likely be content-neutral

on its face. In Universal City Studios, Inc., the court held that the challenged government

regulations were content-neutral because they “target[ed] only the nonspeech component” of the

prohibited actions188 by focusing on the “functional” aspects of computer code that operate

without any human involvement.189 It is possible, though, that a law targeting speech would

nonetheless be content-neutral. For example, the Court has said that “a prohibition against the use

180 Id. at 239–40.

181 Id. at 240.

182 E.g. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2231 (2015) (outlining test for content neutrality).

183 Id.

184 Citizens United v. FEC, 558 U.S. 310, 340 (2010) (quoting FEC v. Wisconsin Right to Life, Inc., 551 U.S. 448, 464

(2007)).

185 Ward v. Rock Against Racism, 491 U.S. 781, 791 (1989) (quoting Clark v. Cmty. for Creative Non-Violence, 468

U.S. 288, 293 (1984)). Ward and some subsequent cases emphasized governmental purpose as a key element in the

inquiry into content neutrality, asking whether the government “adopted a regulation of speech because of

disagreement with the message it conveys.” Id.; see also, e.g., Hill v. Colorado, 530 U.S. 703, 719 (2000). In Reed v.

Town of Gilbert, the Supreme Court said that “the crucial first step in the content-neutrality analysis” is “determining

whether the law is content neutral on its face.” 135 S. Ct. at 2228.

186 273 F.3d 429, 449, 453–58 (2d Cir. 2001).

187 See, e.g., Holder v. Humanitarian Law Project, 561 U.S. 1, 27 (2010).

188 273 F.3d at 454 (saying that a regulation “target[ing] only the nonspeech component” of a particular action is

“content-neutral, just as would be a restriction on trafficking in skeleton keys identified because of their capacity to

unlock jail cells, even though some of the keys happened to bear a slogan or other legend that qualified as a speech

component”).

189 Id. at 451. See also id. at 454 (“Neither the DMCA nor the posting prohibition is concerned with whatever capacity

DeCSS might have for conveying information to a human being, and that capacity, as previously explained, is what

arguably creates a speech component of the decryption code. The DMCA and the posting prohibition are applied to

DeCSS solely because of its capacity to instruct a computer to decrypt CSS.”).

Congressional Research Service

19

Free Speech and the Regulation of Social Media Content

of sound trucks emitting ‘loud and raucous’ noise in residential neighborhoods is permissible if it

applies equally to music, political speech, and advertising.”190

A court might also look to the particular nature of the medium being regulated, asking whether

there are special characteristics that might justify greater regulation.191 The Supreme Court has

said that “[e]ach medium of expression . . . must be assessed for First Amendment purposes by

standards suited to it, for each may present its own problems.”192 The Court has been willing to

extend First Amendment protections that historically applied to speech communicated in

traditional public forums such as streets and sidewalks193 to new mediums for communication,

including video games194 and the internet.195 But the Court has also recognized that the principles

developed “in the context of streets and parks . . . should not be extended in a mechanical way to

the very different context of” newer media.196 While the Court has characterized social media as

“the modern public square,”197 it has not fully clarified what standards should apply to

government regulation of that medium—particularly with respect to social media platforms’ roles

as hosts for others’ speech.

The Supreme Court said in Reno v. ACLU that when considering government regulation of “the

Internet” in general, factors that had previously justified greater regulation of other media did not

apply.198 In that case, the Court held unconstitutional two provisions of the CDA that criminalized

the transmission of certain “indecent” or “patently offensive” material to minors over the

internet.199 The Court rejected the government’s argument that the regulation was permissible

because the internet is analogous to broadcast media, where the Court has permitted greater

regulation of speech.200 The Court noted that unlike the broadcast industry, “the vast democratic

fora of the Internet” had not traditionally “been subject to the type of government supervision and

regulation that has attended the broadcast industry,” and said that “the Internet is not as ‘invasive’

as radio or television.”201 Accordingly, the Court stated that there was “no basis for qualifying the

level of First Amendment scrutiny that should be applied to this medium.”202 However, as will be

190 Cincinnati v. Discovery Network, 507 U.S. 410, 428 (quoting Kovacs v. Cooper, 336 U.S. 77, 87 (1949)).

191 See, e.g., Red Lion Broad. Co. v. FCC, 395 U.S. 367, 386 (1969) (“[D]ifferences in the characteristics of new media

justify differences in the First Amendment standards applied to them.”).

192 Se. Promotions, Ltd. v. Conrad, 420 U.S. 546, 557 (1975).

193 See, e.g., McCullen v. Coakley, 134 S. Ct. 2518, 2529 (2014) (discussing traditional public forums).

194 Brown v. Entm’t Merchs. Ass’n, 564 U.S. 786, 790 (2011) (“[V]ideo games qualify for First Amendment

protection. The Free Speech Clause exists principally to protect discourse on public matters, but we have long

recognized that it is difficult to distinguish politics from entertainment, and dangerous to try.”).

195 E.g., Packingham v. North Carolina, 137 S. Ct. 1730, 1735 (2017) (recognizing “cyberspace” as “the most important

place[] . . . for the exchange of views”). As lower courts have recognized, however, this dicta from the Court does not

necessarily mean that “all conceivable means of communication associated with the Internet necessarily constitute

‘fora’ for protected speech.” Nat’l A-1 Advert. v. Network Sols., Inc., 121 F. Supp. 2d 156, 171 (D.N.H. 2000); see

also, e.g., Hargis v. Bevin, 298 F. Supp. 3d 1003, 1010–12 (E.D. Ky. 2018) (holding that state governor’s social media

pages are not public fora because the governor is speaking personally, on his own behalf, and has not opened the pages

for others’ speech).

196 Ark. Educ. Television Comm’n v. Forbes, 523 U.S. 666, 672–73 (1998).

197 Packingham v. North Carolina, 137 S. Ct. 1730, 1737 (2017).

198 Reno v. ACLU, 521 U.S. 844, 870 (1997) (holding that the factors that justified “qualifying the level of First

Amendment scrutiny that should be applied to” broadcast media did not apply to the Internet”).

199 Id. at 885.

200 Id. at 867, 868–70.

201 Id. at 868–69.

202 Id. at 870.

Congressional Research Service

20

Free Speech and the Regulation of Social Media Content

discussed in more detail below, some scholars have argued that Reno, decided in 1997, does not

specifically address government regulation of modern social media sites, which may present

unique concerns from those discussed in Reno.203

Social Media Sites: Providing a Digital Public Square

Social media sites provide platforms for content originally generated by users. In that capacity,

social media sites decide whether to host users’ content and how that content is presented, and

may alter that content in the process.204 Whether these editorial functions are “speech” protected

by the First Amendment presents an especially difficult question. As one federal appellate court

noted, “entities that serve as conduits for speech produced by others” may “receive First

Amendment protection” if they “engage in editorial discretion” when “selecting which speech to

transmit.”205 On the other hand, the court said, such an entity might not be “a First Amendment

speaker”206 if it indiscriminately and neutrally transmits “any and all users’ speech.”207

Some have argued that social media sites’ publication decisions are protected under the First

Amendment.208 Until recently, academic debate focused largely on whether the algorithms

employed by search engines to retrieve and present results are properly characterized as the

speech of those search engines.209 One scholar argued that search engines’ publication activities

meet at least one of the criteria necessary to qualify for First Amendment protection: these sites

are publishing “sendable and receivable substantive message[s]”—or, in other words, they are

communicating content.210 Another scholar countered this argument by saying that indexing

search results is not equivalent to communicating protected ideas, arguing that to be entitled to

First Amendment protections, content must be “adopted or selected by the speaker as its own.”211

203 See infra notes 307 to 312 and accompanying text.

204 Cf., e.g., Eugene Volokh & Donald M. Falk, First Amendment Protection for Search Engine Results, 8 J.L. ECON. &

POL’Y 883, 884 (2012).

205 U.S. Telecom Ass’n v. FCC, 825 F.3d 674, 742 (D.C. Cir. 2016).

206 Id. at 743.

207 Id. at 742. See also, e.g., Rest. Law Ctr. v. New York City, No. 17 Civ. 9128 (PGG), 2019 U.S. Dist. LEXIS 19378,

at *25 (S.D.N.Y. Feb. 6, 2019) (“Under the First Amendment, an entity’s mere transmission of others’ speech does not

necessarily constitute speech of that entity.”).

208 See, e.g., Eric Goldman, Of Course the First Amendment Protects Google and Facebook (And It’s Not a Close

Question), KNIGHT FIRST AMENDMENT INST. AT COLUMBIA UNIV. (Feb. 2018), https://bit.ly/2R6Ii12 [hereinafter

Goldman, First Amendment Protects Google and Facebook].

209 Compare, e.g., Stuart Minor Benjamin, Algorithms and Speech, 161 U. PA. L. REV. 1445, 1447 (2013) (“[I]f we

accept Supreme Court jurisprudence, the First Amendment encompasses a great swath of algorithm-based decisions—

specifically, algorithm-based outputs that entail a substantive communication.”), and Volokh & Falk, supra note 204, at

884 (arguing that “search engines are speakers”), with Oren Bracha & Frank Pasquale, Federal Search Commission?

Access, Fairness, and Accountability In the Law of Search, 93 CORNELL L. REV. 1149, 1193 (2008) (“It is highly

questionable that search results constitute the kind of speech recognized to be within the ambit of the First Amendment

according to either existing doctrine or any of the common normative theories in the field.”), and Tim Wu, Machine

Speech, 161 U. PA. L. REV. 1495, 1528–30 (2013) (arguing that search results are not speech protected by the First

Amendment).

210 Benjamin, supra note 209, at 1461. Professor Benjamin argues only that this element is necessary, and

acknowledges that it likely is not sufficient for First Amendment protections, asking instead “whether applying the

criteria identified above plus the exceptions the Court has articulated would be inconsistent with some elements of the

Court’s jurisprudence.” Id. at 1462.

211 Wu, supra note 209, at 1529–30.

Congressional Research Service

21

Free Speech and the Regulation of Social Media Content

There are not many court decisions evaluating whether a social media site, by virtue of reprinting,

organizing, or even editing protected speech, is itself exercising free speech rights. While a few

federal courts have held that search engine results212 and decisions about whether to run

advertisements213 are speech protected by the First Amendment, these decisions are, so far,

limited to trial courts and therefore not precedential beyond the facts of those cases. This relative

dearth of cases is likely due in large part to the fact that, as discussed above,214 Section 230 of the

CDA bars a significant number of lawsuits that seek to hold social media providers liable for

publishing others’ content,215 often making it unnecessary to consider whether the First

Amendment protects these publication decisions.216 Section 230 has sometimes been described as

an attempt to protect the freedom of speech on the internet,217 suggesting that its displacement of

the First Amendment is an implicit consequence of Section 230’s speech-protective nature. In

other words, Section 230 creates immunity even where the First Amendment might not.218

Due to the lack of case law examining the issue, commentators have largely analyzed the question

of whether a social media site’s publication decisions are protected by the First Amendment by

analogy to other types of First Amendment cases. At least one scholar has argued that there are

three possible frameworks a court could apply to analyze governmental restrictions on social

media sites’ ability to moderate user content.219 The first analogy would treat social media sites as

equivalent to company towns.220 Under this scenario, social media sites would be treated as state

actors who are themselves bound to follow the First Amendment when they regulate protected

speech. The second possible framework would view social media sites as analogous to special

industries like common carriers or broadcast media, in which the Court has historically allowed

greater regulation of the industries’ speech in light of the need to protect public access for users

212 See e-ventures Worldwide, LLC v. Google, Inc., No. 2:14-cv-646-FtM-PAM-CM, 2017 U.S. Dist. LEXIS 88650, at

*11–12 (M.D. Fla. Feb. 8, 2017); Zhang v. Baidu.com, Inc., 10 F. Supp. 3d 433, 439 (S.D.N.Y. 2014); Search King,

Inc. v. Google Tech., Inc., No. CIV-02-1457-M, 2003 U.S. Dist. LEXIS 27193, at *12 (W.D. Okla. May 27, 2003).

213 Langdon v. Google, Inc., 474 F. Supp. 2d 622, 630 (D. Del. 2007).

214

Supra “Section 230 of the CDA.”

215 47 U.S.C. § 230.

216 See, e.g., Google LLC v. Equustek Sols. Inc., No. 5:17-cv-04207-EJD, 2017 U.S. Dist. LEXIS 182194, at *7 n.2

(N.D. Cal. Nov. 2, 2017) (holding that it was unnecessary to reach the defendant’s claim that its activity was protected

by the First Amendment because the defendant was entitled to immunity under Section 230); Ramey v. Darkside

Prods., No. 02-730, 2004 U.S. Dist. LEXIS 10107, at *12 (D.D.C. May 17, 2004) (same).

217 E.g., Zeran v. Am. Online, Inc. 129 F.3d 327, 330–31 (4th Cir. 1997).

218 Cf., e.g., Gucci Am., Inc. v. Hall & Assocs., 135 F. Supp. 2d 409, 422 (S.D.N.Y. 2001) (“Section 230 reflects a

‘policy choice,’ not a First Amendment imperative, to immunize ISPs from defamation and other ‘tort-based lawsuits,’

driven, in part, by free speech concerns.” (quoting Zeran, 129 F.3d at 330–31)). Interestingly, however, Section 230

creates immunity by providing that service providers cannot be treated as a publisher or speaker of content created by

another. E.g., Kimzey v. Yelp! Inc., 836 F.3d 1263, 1268 (9th Cir. 2016); see also id. at 1271 (“Simply put,

proliferation and dissemination of content does not equal creation or development of content.”). By contrast, First

Amendment protection would likely depend on the providers being seen as the publishers of others’ content. See, e.g.,

Goldman, First Amendment Protects Google and Facebook; Volokh & Falk, supra note 204, at 884. Section 230 was

enacted in direct response to a court decision that had concluded that an internet service provider should be considered

a publisher of defamatory statements that a third party had posted on a bulletin board that it hosted, and could therefore

be subject to suit for libel. See S. REP. NO. 104-230, at 194 (1996); 141 CONG. REC. H8469–70 (daily ed. Aug. 4, 1995)

(statement of Rep. Christopher Cox) (discussing Stratton-Oakmont, Inc. v. Prodigy Services Co., No. 31063/94, 1995

N.Y. Misc. LEXIS 229, at *10–11 (N.Y. Sup. Ct. May 26, 1995)). Arguably, this decision suggests that perhaps absent

Section 230 liability, courts would have viewed service providers as publishers. But cf., e.g., Cubby, Inc. v.

CompuServe, Inc., 776 F. Supp. 135, 140–41 (S.D.N.Y. 1991) (holding CompuServe could not be held liable unless “it

knew or had reason to know of the allegedly defamatory . . . statements” posted on its site).

219 Klonick, supra note 32, at 1658.

220 Id.

Congressional Research Service

22

Free Speech and the Regulation of Social Media Content

of their services.221 The third analogy would treat social media sites like news editors, who

generally receive the full protections of the First Amendment when making editorial decisions.222

It is likely that no one analogy can account for all social media platforms, or all activities

performed by those platforms. Some social media platforms may exercise more editorial control

over user-generated content than others, and any given social media company performs a wide

variety of different functions. Consequently, determining which line of case law is most

analogous will likely depend on the particular activity being regulated.

Social Media Sites as Company Towns

As discussed in more detail above,223 although the First Amendment generally applies only to

government action, the Supreme Court has held that in limited, special circumstances, private

actors should be treated as the government and must comply with constitutional standards when

interacting with others.224 The archetypal case is that of the company town: in Marsh v. Alabama,

the Supreme Court held that the residents of a company-owned town—a town that was

functionally identical to any ordinary town, but for the fact of its ownership—were entitled to the

protections of the First Amendment when distributing religious literature on the streets and

sidewalks in that town.225 Courts have largely held that, under existing Supreme Court precedent,

social media providers do not meet the First Amendment’s state action requirement.226

Commentators have argued, however, that dicta in Supreme Court cases may suggest that social

media sites should be treated differently.227 As an initial matter, there is language in Marsh

suggesting that privately owned property may be subject to the First Amendment if it is opened

for public use:

Ownership does not always mean absolute dominion. The more an owner, for his

advantage, opens up his property for use by the public in general, the more do his rights

become circumscribed by the statutory and constitutional rights of those who use it. Thus,

the owners of privately held bridges, ferries, turnpikes and railroads may not operate them

as freely as a farmer does his farm. Since these facilities are built and operated primarily

to benefit the public and since their operation is essentially a public function, it is subject

to state regulation.228

At least one scholar has argued that, with respect to online forums, “Marsh should be expanded

and read functionally.”229 He suggests that courts should ask whether a given online space is the

221 Id. at 1660.

222 Id. at 1659–60.

223 Supra “First Amendment: State Action Requirement.”

224 See, e.g., Brentwood Acad. v. Tenn. Secondary Sch. Athletic Ass’n, 531 U.S. 288, 298 (2001).

225 326 U.S. 501, 507–09 (1946).

226 See supra note 63.

227 See Klonick, supra note 32, at 1659; Peters, supra note 18, at 1023; Heather Whitney, Search Engines, Social

Media, and the Editorial Analogy, KNIGHT FIRST AMENDMENT INST. AT COLUMBIA UNIV. (Feb. 2018),

https://bit.ly/2Tbl1YP.

228 Marsh, 326 U.S. at 506 (citation omitted). See also, e.g., Evans v. Newton, 382 U.S. 296, 302 (1966) (“A park . . .

traditionally serves the community. Mass recreation through the use of parks is plainly in the public domain; and state

courts that aid private parties to perform that public function on a segregated basis implicate the State in conduct

proscribed by the Fourteenth Amendment. Like the streets of the company town in Marsh v. Alabama . . . , the

predominant character and purpose of this park are municipal.” (citations omitted)).

229 Peters, supra note 18, at 1023.

Congressional Research Service

23

Free Speech and the Regulation of Social Media Content

“functional equivalent” of a traditional public forum230 and should engage in a First Amendment

analysis that treats private ownership as “one factor” when balancing “the autonomy rights of

property owners against the expressive rights of property users.”231 But courts, by and large, have

rejected the broader implications of this language in Marsh,232 and the Supreme Court has held

that the mere fact that a private space is open to the public is not sufficient to “entitle” the public

to the protections of the First Amendment in that space.233

Another scholar, however, has argued that notwithstanding “this more narrow conception of the

public function exception,” social media sites should still be treated as equivalent to the state

under Marsh.234 He claims that social media sites perform a “public function”235 under Marsh by

“providing a space that has the primary purpose of serving as a forum for public communication

and expression, that is designated for that purpose, and that is completely open to the public at

large.”236 In his view, “[s]ince managing public squares and meeting places is something that has

traditionally been done by the government, social network websites therefore serve a public

function that has traditionally been the province of the state.”237 As mentioned above, however,

trial courts have declined to extend Marsh to social media sites, disagreeing that the provision of

a public forum238 or “the dissemination of news and fostering of debate”239 are public functions

that were traditionally and exclusively performed by the government.240

Others have argued that a more recent Supreme Court decision, Packingham v. North Carolina,241

might “signal a shift” in the state action analysis.242 In Packingham, the Court struck down a

North Carolina law that prohibited a registered sex offender from accessing any “commercial

social networking Web site where the sex offender knows that the site permits minor children to

become members or to create or maintain personal Web pages.”243 Critically, the Court stated that

“cyberspace” is today “the most important place[] . . . for the exchange of views” protected by the

First Amendment, analogizing Facebook, LinkedIn, and Twitter to traditional public forums244

and characterizing social media sites as “the modern public square.”245 In light of the importance

230 Id. He suggests that “several considerations would guide that assessment: (1) the nature of the private property

interests at issue, and (2) whether the space is operated for general use by the public for expressive purposes, or

whether the operation is itself a public function, either of which would favor a finding of state action.” Id. at 1024.

231 Id. at 1023.

232 See, e.g., Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 U.S. Dist. LEXIS 51000, at *25–26 (N.D.

Cal. Mar. 26, 2018) (noting but rejecting the “broader language” in Marsh “that could be construed to support

Plaintiff’s position that because Defendants hold out and operate their private property (YouTube) as a forum dedicated

to allowing its users to express diverse points of view, Defendants should be treated as state actors”).

233 Lloyd Corp. v. Tanner, 407 U.S. 551, 569 (1972).

234 Benjamin F. Jackson, Censorship and Freedom of Expression in the Age of Facebook, 44 N.M. L. REV. 121, 146

(2014).

235 Marsh, 326 U.S. at 506.

236 Jackson, supra note 234, at 146.

237 Id.

238 Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 U.S. Dist. LEXIS 51000, at *26 (N.D. Cal. Mar. 26,

2018).

239 Quigley v. Yelp, Inc., No. 17-cv-03771-RS, 2017 U.S. Dist. LEXIS 103771, at *4 (N.D. Cal. July 5, 2017).

240 See Jackson v. Metro. Edison Co., 419 U.S. 345, 352 (1974).

241 137 S. Ct. 1730, 1735 (2017).

242 Klonick, supra note 32, at 1659. See also Whitney, supra note 227.

243 137 S. Ct. at 1733 (quoting N.C. Gen. Stat. Ann. §§ 14-202.5(a), (e) (2015)) (internal quotation mark omitted).

244 Id. at 1735.

245 Id. at 1737 (“Social media allows users to gain access to information and communicate with one another about it on

Congressional Research Service

24

Free Speech and the Regulation of Social Media Content

of these forums, the Court concluded that the statute was too broad and not sufficiently tailored to

serve the government’s asserted interest.246

Some have suggested that, if the Court views social media as “the modern public square,”247 it

may be more willing to say that social media companies “count as state actors for First

Amendment purposes.”248 Indeed, Justice Alito declined to join the majority opinion in

Packingham because he was concerned about the scope of the Court’s “musings that seem to

equate the entirety of the internet with public streets and parks.”249 He argued that this broader

language was “bound to be interpreted by some”—erroneously, in his view—as limiting the

government’s ability to “restrict . . . dangerous sexual predators” from some activities online.250

At least one court, however, has rejected some of the broader implications of this case, noting that

“Packingham did not, and had no occasion to, address whether private social media corporations

like YouTube are state actors that must regulate the content of their websites according to the

strictures of the First Amendment. Instead, . . . Packingham concerned whether North Carolina

ran afoul of the First Amendment . . . .”251

If social media sites were treated as state actors under the First Amendment, then the Constitution

itself would constrain their conduct when they act to restrict users’ protected speech.252 Under this

framework, Congress could enact legislation to remedy violations of free speech rights by social

media entities.253 For instance, Title III of the Civil Rights Act of 1964 authorizes the Attorney

General to bring a civil action against governmental facilities that deny a person equal access “on

account of his race, color, religion, or national origin,”254 essentially granting the Attorney

General the power to sue to enjoin certain acts that would violate the Fourteenth Amendment’s

Equal Protection Clause.255 To take another example, 42 U.S.C. § 1983 allows any person who

has been deprived by a state actor “of any rights, privileges, or immunities secured by the

Constitution” to bring certain civil actions to vindicate those rights in court.256

By contrast, commentators have argued that under this framework, the problems associated with

social media sites hosting too much speech—that is, problems caused by the dissemination of

any subject that might come to mind. By prohibiting sex offenders from using those websites, North Carolina with one

broad stroke bars access to what for many are 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.” (citation omitted)).

246 Id.

247 Id.

248 Whitney, supra note 227.

249 Packingham, 137 S. Ct. at 1738 (Alito, J., concurring).

250 Id.

251 Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 U.S. Dist. LEXIS 51000, at *24 (N.D. Cal. Mar. 26,

2018).

252

See, e.g., Marsh v. Alabama, 326 U.S. 501, 509 (1946).

253 Cf., e.g., City of Boerne v. Flores, 521 U.S. 507, 512 (1997) (“Legislation which deters or remedies constitutional

violations can fall within the sweep of Congress’ enforcement power even if in the process it prohibits conduct which is

not itself unconstitutional and intrudes into ‘legislative spheres of autonomy previously reserved to the States.’”

(quoting Fitzpatrick v. Bitzer, 427 U.S. 445, 455 (1976))).

254 42 U.S.C. § 2000b.

255 See, e.g., United States v. City of Philadelphia, 644 F.2d 187, 197 (3d Cir. 1980).

256 See, e.g., Lozman v. City of Riviera Beach, 138 S. Ct. 1945, 1951 (2018) (considering whether plaintiff sufficiently

alleged a “First Amendment claim for retaliatory arrest” under Section 1983).

Congressional Research Service

25

Free Speech and the Regulation of Social Media Content

things like misinformation and hate speech—would be exacerbated.257 If these companies were

considered equivalent to state actors and their sites were seen as equivalent to traditional public

forums, their ability to regulate speech would be relatively circumscribed.258 And in turn, so

would the government be limited in its ability to require these platforms to take down certain

types of content, if that content qualified as protected speech.259 Thus, one scholar predicted that

under this framework, “[a]ll but the very basest speech would be explicitly allowed and

protected—making current problems of online hate speech, bullying, and terrorism, with which

many activists and scholars are concerned, unimaginably worse.”260

However, to the extent that a federal regulation infringed on speech properly attributed to the

social media sites, rather than their users—and this speech could include not only content

originally generated by the social media companies, but also their editorial decisions about usergenerated content—it could implicate an open First Amendment question. State and local

governments are constrained by the First Amendment when they interact with individuals, but the

Supreme Court has never squarely resolved whether states and municipalities could themselves

assert First Amendment rights against the federal government.261 At least one scholar has argued

that the First Amendment should protect government speech in certain circumstances.262 And in a

2015 case, the Supreme Court said that a private party could not force a state to include certain

messages in its own speech, suggesting that governments do have some right to speak for

themselves.263 On the other hand, Justice Stewart argued in a 1973 concurring opinion that the

government has no First Amendment rights, significantly, maintaining that the Court should not

treat broadcasters as state actors because it would “simply strip” them of their First Amendment

rights.264 Lower courts have largely followed Justice Stewart’s view and assumed that state actors

257 E.g., Keller, supra note 76, at 13 (“[H]old[ing] platforms to the same rules as the government . . . . would . . . require

platforms to preserve speech that many people find obnoxious, immoral, or dangerous.”); Klonick, supra note 32, at

1659 (“Interpreting online platforms as state actors . . . would also likely create an internet nobody wants. Platforms

would no longer be able to remove obscene or violent content.”).

258 While the category of “traditional public forums” has been interpreted relatively narrowly, the “same standards” that

apply there apply also in “designated public forums,” spaces that “the government has ‘intentionally opened up for

[the] purpose’” of providing a public forum. Minn. Voters All. v. Mansky, 138 S. Ct. 1876, 1885 (2018). Lower courts

have concluded that at least some government sites on social media qualify as public forums. See, e .g., Davison v.

Randall, 912 F.3d 666, 682 (4th Cir. 2019).

259 Cf., e.g., Davison, 912 F.3d at 687 (concluding that local government official engaged in unconstitutional viewpoint

discrimination when she banned a private party for posting a comment alleging governmental corruption).

260 Klonick, supra note 32, at 1659.

261 Eugene Volokh, Do State and Local Governments Have Free Speech Rights?, WASH. POST. (June 24, 2015),

https://wapo.st/2HJRQt8; see also United States v. Am. Library Ass’n, 539 U.S. 194, 211 (2003) (stating that the Court

“need not decide” whether public libraries could assert First Amendment rights against a federal regulation).

262 David Fagundes, State Actors as First Amendment Speakers, 100 NW. U. L. REV. 1637, 1640 (2006) (arguing that

government speech should be protected “where the expressive conduct at issue is constitutive of the public function of

the entity speaking, so that restricting expression would rob the speaker of a core purpose for which it was created” and

if “the government speech at issue furthers the First Amendment value of democratic self-government”).

263 Walker v. Tex. Div., Sons of Confederate Veterans, Inc., 135 S. Ct. 2239, 2253 (2015) (“[J]ust as Texas cannot

require [the Sons of Confederate Veterans (SCV)] to convey ‘the State’s ideological message,’ SCV cannot force Texas

to include a Confederate battle flag on its specialty license plates.” (quoting Wooley v. Maynard, 430 U.S. 705, 715

(1977))). It can be difficult, however, to discern government speech from private speech. See, e.g., Matal v. Tam, 137

S. Ct. 1744, 1757–58 (2017); Walker, 135 S. Ct. at 2245, 2250.

264 Columbia Broad. Sys., Inc. v. Democratic Nat’l Comm., 412 U.S. 94, 139 (1973) (Stewart, J., concurring) (“The

First Amendment protects the press from governmental interference; it confers no analogous protection on the

Government. To hold that broadcaster action is governmental action would thus simply strip broadcasters of their own

First Amendment rights.”).

Congressional Research Service

26

Free Speech and the Regulation of Social Media Content

may not claim the protection of the First Amendment.265 Accordingly, it is possible that treating

social media sites like state actors would “strip [them] of their own First Amendment rights.”266

Social Media Sites as Broadcasters or Cable Providers

Alternatively, courts could analogize social media sites to certain industries, like broadcast media,

where the Supreme Court has traditionally allowed greater regulation of protected speech.267

These cases have their roots in the common law doctrines related to common carriers.268

Historically, a common carrier is an entity that “holds itself out to the public as offering to

transport freight or passengers for a fee.”269 Often, these companies received government licenses

authorizing their operations.270 Common carriers have traditionally been subject to heightened

legal duties and generally could not refuse paying customers.271 Some of these common law

doctrines have been incorporated into modern regulation of communications industries: federal

statutes treat providers of telecommunications services as common carriers that are subject to

certain requirements,272 and authorize the regulation of radio and television broadcasters.273 While

acknowledging that these companies are private entities who do retain First Amendment rights,

the Supreme Court has nonetheless allowed some regulation of these rights, in light of the

heightened government interests in regulating such entities.274 As one federal appellate court has

put it, the general “absence of any First Amendment concern” with “equal access obligations” in

this area “rests on the understanding that such entities, insofar as they are subject to equal access

mandates, merely facilitate the transmission of the speech of others rather than engage in speech

in their own right.”275 However, courts have not treated all entities equated to common carriers

identically.276

265 See Fagundes, supra note 262, at 1643; cf. id. (arguing that these cases are “not as unanimous as it has often been

described”).

266 Columbia Broad. Sys., Inc., 412 U.S. at 139. Cf., e.g., Tex. Div., Sons of Confederate Veterans, Inc., 135 S. Ct. at

2255 (Alito, J., dissenting) (“Unfortunately, the Court’s decision categorizes private speech as government speech and

thus strips it of all First Amendment protection.”)

267 See Klonick, supra note 32, at 1612; Genevieve Lakier, The Problem Isn’t the Use of Analogies but the Analogies

Courts Use, KNIGHT FIRST AMENDMENT INST. AT COLUMBIA UNIV. (Feb. 2018), https://bit.ly/2Mw9fpo.

268 Cf., e.g., Klonick, supra note 32, at 1660; Klint Finley, Former FCC Chair Tom Wheeler Says the Internet Needs

Regulation, WIRED (Feb. 27, 2019, 7:00 AM), https://bit.ly/2tE8ulP.

269 BLACK’S LAW DICTIONARY (10th ed. 2014).

270 See, e.g., Landstar Express Am., Inc. v. Fed. Mar. Comm’n, 569 F.3d 493, 494 (D.C. Cir. 2009); TRT Telecomms.

Corp. v. FCC, 876 F.2d 134, 137 (D.C. Cir. 1989); Am. Trucking Ass’ns v. Interstate Commerce Comm’n, 770 F.2d

535, 543 (5th Cir. 1985).

271 BLACK’S LAW DICTIONARY (10th ed. 2014). See also, e.g., Barbara A. Cherry, Utilizing “Essentiality of Access”

Analyses to Mitigate Risky, Costly and Untimely Government Interventions in Converging Telecommunications

Technologies and Markets, 11 COMMLAW CONSPECTUS 251, 256–57 (2003).

272 See, e.g., Direct Commc’ns Cedar Valley, LLC v. FCC, 753 F.3d 1015, 1094 (10th Cir. 2014) (discussing Title II of

the Telecommunications Act of 1996, 47 U.S.C. § 153).

273 See, e.g., Red Lion Broad. Co. v. FCC, 395 U.S. 367, 369–70 (1969) (discussing the Communications Act of 1934,

47 U.S.C. § 301).

274 See, e.g., Reno v. ACLU, 521 U.S. 844, 868 (1997) (reviewing some of the “special justifications for regulation of

the broadcast media that are not applicable to other speakers”).

275 U.S. Telecom Ass’n v. FCC, 825 F.3d 674, 741 (D.C. Cir. 2016).

276 Cf., e.g., Daniel T. Deacon, Common Carrier Essentialism and the Emerging Common Law of Internet Regulation,

67 ADMIN. L. REV. 133, 134–35 (2015) (arguing that “common carrier essentialism provides a fundamentally unstable

framework for the Commission to develop Internet policy” in part because “it is far from clear what comprises the

essence of a common carrier”).

Congressional Research Service

27

Free Speech and the Regulation of Social Media Content

Broadcasters

In Red Lion Broadcasting Co. v. FCC, the Supreme Court approved of a specific application of

the Federal Communication Commission’s (FCC’s) “fairness doctrine.”277 The FCC rule

challenged in Red Lion required broadcasters to give political candidates a reasonable opportunity

to respond to any personal attacks published by the broadcaster or to any editorials in which a

broadcaster endorsed or opposed particular candidates.278 The broadcasters argued that these

regulations violated the First Amendment, abridging “their freedom of speech and press” by

preventing them from “exclud[ing] whomever they choose” from their allotted frequencies.279

The Supreme Court noted the unique nature of the broadcast industry, stating that due to “the

scarcity of radio frequencies,”280 “it is idle to posit an unabridgeable First Amendment right to

broadcast comparable to the right of every individual to speak, write, or publish.”281 This is why,

the Court said, it had previously allowed regulations of broadcast media—namely, a licensing

system—that might otherwise violate the First Amendment.282 The Court emphasized that “[i]t is

the right of the viewers and listeners, not the right of the broadcasters, which is paramount,”

highlighting “the right of the public to receive suitable access to social, political, esthetic, moral,

and other ideas and experiences.”283 Ultimately, the Court held that “[i]n view of the scarcity of

broadcast frequencies, the Government’s role in allocating those frequencies, and the legitimate

claims of those unable without governmental assistance to gain access to those frequencies for

expression of their views,” the challenged regulations were constitutional.284

In subsequent cases, the Supreme Court has reaffirmed that “of all forms of communication, it is

broadcasting that has received the most limited First Amendment protection.”285 The Court has

recognized that broadcasters do engage in speech activity protected by the First Amendment,

most notably when a broadcaster “exercises editorial discretion in the selection and presentation

of its programming.”286 The Court has said that “[a]lthough programming decisions often involve

the compilation of the speech of third parties, the decisions nonetheless constitute communicative

acts.”287 Notwithstanding this conclusion, however, the Court has said that in this area, when

evaluating broadcasters’ First Amendment claims, it will “afford great weight to the decisions of

Congress and the experience of the [FCC].”288

277 See 395 U.S. at 375.

278 Id. at 373–75.

279 Id. at 386.

280 Id. at 390.

281 Id. at 388.

282 Id. at 389. Ordinarily, licensing systems applied to speech might raise First Amendment concerns under the doctrine

generally disfavoring prior restraints. See, e.g., Se. Promotions, Ltd. v. Conrad, 420 U.S. 546 (1975). But in Red Lion,

the Court confirmed that denying a broadcast “license because ‘the public interest’ requires it ‘is not a denial of free

speech.’” Red Lion, 395 U.S. at 389 (quoting Nat’l Broad. Co. v. United States, 319 U.S. 190, 227 (1943)).

283 Red Lion, 395 U.S. at 390.

284 Id. at 400–01.

285 FCC v. Pacifica Found., 438 U.S. 726, 793 (1978).

286 Ark. Educ. TV Comm’n v. Forbes, 523 U.S. 666, 674 (1998).

287 Id.

288 Columbia Broad. Sys., Inc. v. Democratic Nat’l Comm., 412 U.S. 94, 103 (1973).

Congressional Research Service

28

Free Speech and the Regulation of Social Media Content

Cable Television

Significantly, the Supreme Court has declined to extend this special deference to government

regulation of broadcasters to other forms of media. For example, in Turner Broadcasting Systems

v. FCC, the Court concluded that the “less rigorous” First Amendment scrutiny that applies to

broadcast regulation should not be extended to the “regulation of cable television.”289 The Court

was considering the FCC’s “must-carry” regulations, which required cable television broadcasters

to set aside a portion of their channels for the transmission of local broadcast television

stations.290 The Court said that cable television “does not suffer from the inherent limitations,” in

terms of the scarcity of frequencies, “that characterize the broadcast medium,” consequently

concluding that the “unique physical characteristics of cable transmission . . . . do not require the

alteration of settled principles of our First Amendment jurisprudence.”291 Accordingly, the Court

has subjected laws that restrict cable providers’ protected speech to greater scrutiny than

restrictions on broadcast media.292

But the Court noted in a subsequent decision that “[c]able television, like broadcast media,

presents unique problems . . . which may justify restrictions that would be unacceptable in other

contexts.”293 And in Turner Broadcasting itself, the Court did cite “special characteristics of the

cable medium” to justify applying a lower level of scrutiny.294 The Court recognized that

“[r]egulations that discriminate among media, or among different speakers within a single

medium, often present serious First Amendment concerns” that trigger strict scrutiny.295 But

notwithstanding this general rule, the Court explained that “heightened scrutiny is unwarranted

where,” as with the must-carry provisions, the “differential treatment is ‘justified by some special

characteristic of’ the particular medium being regulated.”296 Courts have sometimes interpreted

Turner Broadcasting to mean that at least certain types of regulations on cable television will

receive less scrutiny than, for example, a regulation affecting speech in a traditional public

forum.297

289 512 U.S. 622, 638 (1994).

290 Id. at 630–32.

291 Id. at 639.

292 Compare, e.g., FCC v. Pacifica Found., 438 U.S. 726, 745–51 (1978) (upholding FCC sanction of broadcast

containing “patently offensive words” even though it targeted constitutionally protected speech, because broadcasting’s

pervasiveness and its availability to children “justify special treatment”), with, e.g., United States v. Playboy Entm’t

Group, 529 U.S. 803, 809, 827 (2000) (holding that federal statute restricting the availability of “sexually explicit

[cable] channel[s]” discriminated on the basis of content and was unconstitutional under a strict scrutiny analysis).

293 Playboy Entm’t Group, 529 U.S. at 813. In particular, the Court noted the possibility of “unwanted, indecent speech

that comes into the home.” Id. at 814. However, the Court nonetheless applied strict scrutiny to the content-based

regulations challenged in that case, distinguishing Pacifica, see supra note 292, based on “a key difference between

cable television and the broadcasting media”: the fact that households can block cable channels. 529 U.S. at 815.

294 512 U.S. at 661.

295 Id. at 659.

296

Id. at 660–61 (quoting Minneapolis Star & Tribune Co. v. Minn. Comm’r of Revenue, 460 U.S. 575, 585 (1983)). In

particular, the Court said that the must-carry provisions were justified by “the bottleneck monopoly power exercised by

cable operators and the dangers this power poses to the viability of broadcast television.” Id. at 661. Further, the Court

said that because the regulations were “broad-based” and widely applicable to “almost all cable systems,” they were

“not structured in a manner that carries the inherent risk of undermining First Amendment interests.” Id.

297 See, e.g., Denver Area Educ. Telecoms. Consortium v. FCC, 518 U.S. 727, 769 (1996) (Stevens, J., concurring)

(noting that Turner Broadcasting was decided “without reference to our public forum precedents,” and arguing that

when regulating cable, the federal government “deserves more deference than a rigid application of the public forum

doctrine would allow”); see also, e.g., Comcast Cable Communs., LLC v. FCC, 717 F.3d 982, 993 (D.C. Cir. 2013)

(Kavanaugh, J., concurring) (stating that the Supreme Court’s ruling that the “must-carry provisions might satisfy

Congressional Research Service

29

Free Speech and the Regulation of Social Media Content

Ultimately, however, the Turner Broadcasting Court cited two justifications for applying

intermediate scrutiny, rather than strict scrutiny, to the FCC’s must-carry provisions, making it

unclear which rationale the Court relied on to uphold the regulations. Prior to its discussion of

cable’s special characteristics, the Court concluded that intermediate scrutiny was appropriate

because the must-carry provisions were “content-neutral restrictions that impose[d] an incidental

burden on speech.”298 The Court noted that while the rules did “interfere with cable operators’

discretion . . . , the extent of the interference [did] not depend upon the content of the cable

operators’ programming.”299 Although the must-carry provisions did “distinguish between

speakers,” that discrimination was “based only upon the manner in which speakers transmit their

messages to viewers, and not upon the messages they carry,” and, therefore, the rules were

content-neutral on their face.300 Thus, it is somewhat unclear to what extent the Court’s decision

to apply intermediate scrutiny in Turner Broadcasting rested on “special characteristics of the

cable medium” and to what extent it depended on a more overarching First Amendment principle

regarding content neutrality.301

Treating Social Media Like Broadcast or Cable

While the Supreme Court has identified “unique problems” that may justify greater regulation of

broadcast and cable,302 it has expressly held that the factors that justify more extensive regulation

of the broadcast media “are not present in cyberspace.”303 In Reno v. ACLU, decided in 1997, the

Court said that the internet had not historically “been subject to the type of government

supervision and regulation that has attended the broadcast industry,” that the internet was not “as

‘invasive’ as radio or television” because a person had to take affirmative action to receive a

particular communication on the internet, and that the internet could “hardly be considered a

‘scarce’ expressive commodity.”304 Consequently, in the Court’s view, the factors that justified

“qualifying the level of First Amendment scrutiny that should be applied to” broadcast media did

not apply to the internet.305 In Reno, the Court ultimately held that two provisions of the CDA that

criminalized speech based on its content were unconstitutionally vague and overbroad.306

Several legal scholars have argued that, contrary to the Court’s conclusion in Reno, the internet is

analogous to traditional broadcast media and therefore should be subject to greater regulation.307

intermediate First Amendment scrutiny . . . rested . . . on special characteristics of the cable medium”) (internal

quotation mark omitted).

298 Turner Broad. Sys., 512 U.S. at 662.

299 Id. at 643–44. Further, the Court noted that the “overriding congressional purpose” was “unrelated to the content of

expression disseminated by cable and broadcast speakers.” Id. at 647.

300 Id. at 645; see also id. (“So long as they are not a subtle means of exercising a content preference, speaker

distinctions of this nature are not presumed invalid under the First Amendment.”).

301 See id. at 653–61. Cf. United States v. Playboy Entm’t Group, 529 U.S. 803, 809, 813–15 (2000) (applying strict

scrutiny to content-based speech restriction regardless of the “unique problems” presented by the cable medium).

302 Playboy Entm’t Group, 529 U.S. at 813.

303 Reno v. ACLU, 521 U.S. 844, 868 (1997).

304 Id. at 868–70.

305 Id. at 870.

306 See id. at 871–72.

307 E.g., Bracha & Pasquale, supra note 209, at 1191; Klonick, supra note 32, at 1660–61. In addition, the D.C. Circuit

upheld an FCC order classifying broadband service as a common carrier against a First Amendment challenge,

concluding that internet broadband providers “act as ‘mere conduits for the messages of others, not as agents exercising

editorial discretion subject to First Amendment protections,’” noting that broadband providers do not, in fact, exercise

significant control over the content that internet users access. 825 F.3d 674, 741 (D.C. Cir. 2016) (quoting In re

Congressional Research Service

30

Free Speech and the Regulation of Social Media Content

Scholars have argued that as the internet has developed, it has “reproduce[d] the traditional

speech-hierarchy of broadcasting”: “small, independent speakers [are] relegated to an

increasingly marginal position while a handful of commercial giants capture the overwhelming

majority of users’ attention and reemerge as the essential gateways for effective speech.”308 Thus,

as one scholar argued, “the hold of certain platforms” over “certain mediums of speech” has

“created scarcity.”309 Further, especially as compared to the internet in the late 1990s, when Reno

was decided, the internet is “now more invasive in everyday life”—arguably more invasive even

than television and radio.310 Another commentator has claimed that rather than traditional

broadcast media, search engines might be more analogous to cable providers.311 In her view,

search engines, “like cable companies,” “provide access to the speech of others” but also

“exercise some degree of editorial discretion over whom they provide access to.”312 The analogy

may be extended to social media sites, as well, because, like search engines, they also exercise

editorial discretion regarding who can post and view content on their sites, and regarding how

user-generated content is presented.

One lower court rejected these arguments, with respect to search engines, in Zhang v. Baidu.com,

Inc.313 In that case, the plaintiffs argued that Baidu, a Chinese search engine, had violated federal

and state civil rights laws by blocking “from its search results . . . information concerning ‘the

Democracy movement in China’ and related topics.”314 Baidu argued that its decisions to block

these search results were protected by the First Amendment.315 The judge noted that “some

scholars” had argued that under Turner Broadcasting, search-engine results should receive a

“lower level of protection.”316 However, in the court’s view, the First Amendment “plainly

shield[ed]” the search engine from this particular lawsuit because the plaintiff’s own suit sought

“to hold Baidu liable for, and thus punish Baidu for, a conscious decision to design its searchengine algorithms to favor certain expression on core political subjects over other expression on

those same political subjects.”317 Accordingly, the court said that “Turner’s three principal

rationales for applying a lower level of scrutiny to the must-carry cable regulations—namely, that

cable companies were mere conduits for the speech of others, that they had the physical ability to

silence other speakers, and that the regulations at issue were content-neutral—[we]re

inapplicable” to the case before it.318 The court concluded that Baidu was acting as more than a

conduit for others’ speech, at least according to the plaintiffs’ allegations, that Baidu lacked “the

Protecting and Promoting the Open Internet, No. 15-24, 30 FCC Rcd. 5601, 5870 (2015)), cert. denied, 139 S. Ct. 475

(2018). Accord, e.g., Jack M. Balkin, Free Speech and Press in the Digital Age: The Future of Free Expression in a

Digital Age, 36 PEPP. L. REV. 427, 430 (2009) (arguing that net neutrality rules do not violate the First Amendment

because they “treat network providers as conduits for the speech of others and regulate them in their capacity as

conduits”). The FCC subsequently reversed its original order. See In re Restoring Internet Freedom, No. 17-166, 33

FCC Rcd. 311 (2018).

308 Bracha & Pasquale, supra note 209, at 1158.

309 Klonick, supra note 32, at 1661.

310 Id.

311

Lakier, supra note 267.

312 Id.

313 10 F. Supp. 3d 433, 439 (S.D.N.Y. 2014).

314 Id. at 434–35.

315 Id. at 440.

316 Id. at 439.

317 Id. at 440.

318 Id.

Congressional Research Service

31

Free Speech and the Regulation of Social Media Content

physical power to silence anyone’s voices,” and that a judicial decision penalizing “Baidu

precisely because of what it does and does not choose to say” would not be content-neutral.319

If courts treated social media sites like broadcast media or like cable providers, they would be

more likely to uphold government regulation of social media providers. As a preliminary inquiry,

a court would likely ask what regulations could be justified by specific characteristics of the

regulated medium.320 If a court believed that the internet in general, or social media in particular,

shared relevant characteristics with either traditional broadcast media or with cable providers,

then it would be more likely to allow the types of regulations that have traditionally been

permitted in those contexts. Thus, a court might ask whether social media sites, like cable

companies, exercise a “bottleneck monopoly power”321 or whether, like broadcast television or

radio, social media platforms suffer from a “scarcity” problem in terms of the number of

platforms for speech or are so “invasive” as to justify regulation to address these problems.322

Related, courts might also ask whether the regulations are intended to increase the amount of

information or expression available to the public.323 Thus, if social media sites present distinct

problems that threaten the use of the medium for communicative or expressive purposes,324 courts

might approve of regulations intended to solve those problems—particularly if those regulations

are content-neutral.325 These same types of considerations would likely apply both to regulations

requiring these platforms to carry certain content and to those requiring the platforms not to carry

certain content.326 But, at least for the time being, without an intervening change in the law, lower

courts seem likely to follow Reno and conclude that there is “no basis for qualifying the level of

First Amendment scrutiny that should be applied to” the internet.327

319 Id. at 441.

320 See Turner Broad. Sys. v. FCC, 512 U.S. 622, 661 (1994); Red Lion Broad. Co. v. FCC, 395 U.S. 367, 386 (1969).

321 Turner Broad. Sys., 512 U.S. at 661.

322 Reno v. ACLU, 521 U.S. 844, 868 (1997). For example, Facebook reportedly collects data about users from other

apps, even if the user of those apps “has no connection to Facebook.” Sam Schechner & Mark Secada, You Give Apps

Sensitive Personal Information. Then They Tell Facebook., WALL STREET J. (Feb. 22, 2019, 11:07 AM),

https://on.wsj.com/2GFskWf. Assuming that this is true, a regulation aimed at the invasiveness of Facebook’s data

collection activities might qualify for review under the lower standards of Red Lion or Turner Broadcasting even if it

does also regulate protected speech. See also, e.g., Kashmir Hill, I Cut the ‘Big Five’ Tech Giants from My Life. It Was

Hell, GIZMODO (Feb. 7, 2019, 12:00 PM), https://bit.ly/2BjWyKb (describing the difficulty of avoiding using Amazon,

Facebook, Google, Microsoft, and Apple).

323 See, e.g., Turner Broad. Sys., 512 U.S. at 663 (“[A]ssuring that the public has access to a multiplicity of information

sources is a governmental purpose of the highest order, for it promotes values central to the First Amendment.”); FCC

v. League of Women Voters, 468 U.S. 364, 380 (1984) (“[A]lthough the broadcasting industry plainly operates under

restraints not imposed upon other media, the thrust of these restrictions has generally been to secure the public’s First

Amendment interest in receiving a balanced presentation of views on diverse matters of public concern.”).

324 See, e.g., Se. Promotions, Ltd. v. Conrad, 420 U.S. 546, 557 (1975) (“Each medium of expression, of course, must

be assessed for First Amendment purposes by standards suited to it, for each may present its own problems.”).

325 Cf., e.g., Turner Broad. Sys. v. FCC, 520 U.S. 180, 213 (1997) (concluding that must-carry provisions “ensure[] that

a number of local broadcasters retain cable carriage, with the concomitant audience access and advertising revenues

needed to support a multiplicity of stations”); see also Ark. Educ. Television Comm’n v. Forbes, 523 U.S. 666, 675

(1998) (saying that while the First Amendment does not “of its own force . . . compel public broadcasters to allow third

parties access to their programming, neither would it necessarily “bar the legislative imposition of neutral rules for

access to public broadcasting”).

326 In the realm of broadcast media, the Supreme Court has applied the same type of analysis when considering, for

example, the fairness doctrine, see Red Lion Broad. Co. v. FCC, 395 U.S. 367, 375 (1969), as when considering FCC

action censoring offensive words, see FCC v. Pacifica Found., 438 U.S. 726, 745–51 (1978).

327 Reno v. ACLU, 521 U.S. 844, 870 (1997).

Congressional Research Service

32

Free Speech and the Regulation of Social Media Content

Social Media Sites as Editors

The third analogy courts might use to analyze whether social media sites moderating user content

are exercising protected speech rights is that of the newspaper editor.328 In Miami Herald

Publishing Co. v. Tornillo, the Supreme Court held that when newspapers “exercise . . . editorial

control and judgment,” such as choosing what “material [will] go into a newspaper,” and making

“decisions . . . as to limitations on the size and content of the paper, and treatment of public issues

and public officials,” they are exercising free speech rights protected by the First Amendment.329

The Court in that case was considering the constitutionality of a state law that gave political

candidates the “right to reply to press criticism” of the candidate.330 A newspaper challenged this

statute, arguing that forcing it to print content that it would not otherwise publish violated the

First Amendment.331 The government argued that its law was necessary due to the fact that

relatively few news outlets exercised essentially a “monopoly” on “the ‘marketplace of ideas.’”332

The regulation, in the state’s view, “[e]nsure[d] fairness and accuracy” and “provide[d] for some

accountability.”333

The Supreme Court unanimously rejected this argument, noting that while “press responsibility”

may be a “desirable goal,” it was “not mandated by the Constitution” and could not “be

legislated.”334 The state law impermissibly “exact[ed] a penalty on the basis of the content of the

newspaper” by forcing newspapers to spend money to print the replies and by “taking up space

that could be devoted to other material.”335 Further, the Court held, “[e]ven if a newspaper would

face no additional costs to comply with a compulsory access law and would not be forced to

forgo publication of news or opinion by the inclusion of a reply,” the law violated the First

Amendment “because of its intrusion into the function of editors.”336 Because newspapers

exercise “editorial control and judgment,” the Court said, they are “more than a passive receptacle

or conduit for news, comment, and advertising,” and instead engage in protected speech.337

The Court has recognized this First Amendment protection for editorial judgments outside the

context of newspapers, stating more generally that “compelling a private corporation to provide a

forum for views other than its own may infringe the corporation’s freedom of speech.”338 For

example, the Supreme Court said in Arkansas Educational Television Commission v. Forbes that

“[w]hen a public broadcaster exercises editorial discretion in the selection and presentation of its

328 E.g., Klonick, supra note 32, at 1612. A related but distinct analogy might treat social media sites like booksellers,

who have the “First Amendment right to distribute and facilitate protected speech.” Publius v. Boyer-Vine, 237 F.

Supp. 3d 997, 1008 (E.D. Cal. 2017) (holding that the owner of a blog had “a First Amendment right to distribute and

facilitate” comments on that blog, even though the blogger “did not produce the content” in the comment).

329 418 U.S. 241, 258 (1974).

330 Id. at 247. See also id. at 244 (describing statute as providing “that if a candidate for nomination or election is

assailed regarding his personal character or official record by any newspaper, the candidate has the right to demand that

the newspaper print, free of cost to the candidate, any reply the candidate may make to the newspaper’s charges”).

331

Id. at 245.

332 Id. at 251.

333 Id.

334 Id. at 256.

335 Id.

336 Id. at 258.

337 Id.

338 Pac. Gas & Elec. Co. v. Pub. Utils. Comm’n, 475 U.S. 1, 9 (1986).

Congressional Research Service

33

Free Speech and the Regulation of Social Media Content

programming, it engages in speech activity.”339 And in Pacific Gas & Electric Co. v. Public

Utilities Commission, the Court recognized that a utility company had a First Amendment interest

in selecting the content contained in its monthly newsletter.340 The Court said in Pacific Gas &

Electric Co. that a state regulatory commission could not require the utility to grant access to

entities who disagreed with the utility’s views.341 This regulation infringed on the utility

company’s First Amendment rights by compelling it “to assist in disseminating the speaker’s

message” and by requiring it “to associate with speech with which [the company] may disagree,”

forcing the company to respond to those arguments.342

To take another example, in Hurley v. Irish-American Gay, Lesbian and Bisexual Group of

Boston, the Court held that the private organizers of a parade had a First Amendment right to

exclude the Irish-American Gay, Lesbian and Bisexual Group of Boston (GLIB) from the

parade.343 GLIB had sued the parade organizers, arguing that their exclusion violated

Massachusetts’s antidiscrimination laws by barring them from a public accommodation on the

basis of sexual orientation, and state courts had agreed that GLIB’s exclusion violated state

law.344 The parade organizers, however, claimed that the parade was an expressive activity and

that forcing them to include GLIB’s speech in the parade violated their First Amendment

rights.345 The Supreme Court held first that a parade did qualify as “protected expression,” even

though most of the speech in the parade was not that of the organizers themselves.346 The Court

said that “a private speaker does not forfeit constitutional protection simply by combining

multifarious voices, or by failing to edit their themes to isolate an exact message as the exclusive

subject matter of the speech.”347 As an example, the Court noted that “[c]able operators . . . are

engaged in protected speech activities even when they only select programming originally

produced by others.”348 Accordingly, the Court concluded that the selection of parade participants

was protected activity under the First Amendment.349

Consequently, in the Hurley Court’s view, characterizing the parade as a public accommodation

under the state’s antidiscrimination law “had the effect of declaring the sponsors’ speech itself to

be the public accommodation,” and this exercise of state power “violate[d] the fundamental rule

of protection under the First Amendment, that a speaker has the autonomy to choose the content

of his own message.”350 GLIB argued that this application of the state’s public accommodation

law should be upheld under Turner Broadcasting,351 claiming that the parade organizers, “like a

cable operator,” were “merely a conduit for the speech of participants in the parade rather than

itself a speaker.”352 The Court disagreed, saying that unlike the cable operators, “GLIB’s

339 523 U.S. 666, 674 (1998).

340 475 U.S. at 8–9.

341 Id. at 19–20 (holding that the commission’s order failed strict scrutiny).

342 Id. at 15–16.

343 515 U.S. 557, 581 (1995).

344 Id. at 561–66.

345

Id.

346 Id. at 569.

347 Id. at 569–70.

348 Id. at 570.

349 Id.

350 Id. at 573.

351 Turner Broad. Sys. v. FCC, 512 U.S. 622, 661 (1994).

352 Hurley, 515 U.S. at 575 (internal quotation marks omitted).

Congressional Research Service

34

Free Speech and the Regulation of Social Media Content

participation would likely be perceived as” a decision of the parade organizers that GLIB’s

“message was worthy of presentation and quite possibly of support as well.”353 The better

analogy, in the Court’s view, was to a newspaper.354 The Court said that viewers understand that

cable programming consists of “individual, unrelated segments that happen to be transmitted

together,” but in contrast, “the parade’s overall message is distilled from the individual

presentations along the way, and each unit’s expression is perceived by spectators as part of the

whole.”355

By contrast, the Supreme Court has rejected the application of Tornillo in cases where compelling

a private entity to grant access to third parties would not affect the entity’s own speech.356 First, in

PruneYard Shopping Center v. Robins, a private shopping center, PruneYard, had “a policy not to

permit any visitor or tenant to engage in any publicly expressive activity,” and pursuant to that

policy, asked a number of students distributing pamphlets and seeking signatures on petitions to

leave.357 In a suit brought by the students, the California Supreme Court held that PruneYard’s

action violated state law, holding that the students “were entitled to conduct their activity on

PruneYard property.”358 PruneYard argued that this decision violated their own free speech rights,

claiming that “a private property owner has a First Amendment right not to be forced by the State

to use his property as a forum for the speech of others.”359 The Court rejected this argument,

noting that the government was not forcing PruneYard itself to espouse any specific views, and

that PruneYard could “expressly disavow any connection with” any particular message.360 The

Court said that under the circumstances, “[t]he views expressed by members of the public” would

“not likely be identified with those of the owner.”361

The Court distinguished Tornillo on similar grounds in Rumsfeld v. Forum for Academic and

Institutional Rights, Inc. (FAIR).362 In that case, a group of law schools represented by FAIR

protested the Solomon Amendment, which specified “that if any part of an institution of higher

education denies military recruiters access equal to that provided other recruiters, the entire

institution would lose certain federal funds.”363 Prior to the passage of the Solomon Amendment,

353 Id.

354 Id.

355 Id. at 576–77.

356 See, e.g., Pac. Gas & Elec. Co. v. Pub. Utils. Comm’n, 475 U.S. 1, 12 (1986) (discussing and distinguishing

PruneYard Shopping Center v. Robins, 447 U.S. 74, 87 (1980), noting that in PruneYard, there was no “concern that

access to this area might affect the shopping center owner’s exercise of his own right to speak: the owner did not even

allege that he objected to the content of the pamphlets; nor was the access right content based”).

357 447 U.S. 74, 77 (1980). The state court had concluded that the California Constitution “protects ‘speech and

petitioning, reasonably exercised, in shopping centers even when the centers are privately owned.’” Id. at 78 (quoting

Robins v. Pruneyard Shopping Ctr. 592 P.2d 341, 347 (Cal. 1979)).

358 Id. at 78.

359 Id. at 85. PruneYard also raised a Fifth Amendment challenge, arguing that by denying them their right to exclude

others from their property, the state had unlawfully taken their property. Id. at 82. The Court disagreed. Id. at 85.

360 Id. at 87. PruneYard cited Tornillo, among other cases, as support for its position. Id. at 87–88. The Supreme Court

rejected this argument rather tersely, saying that Tornillo was based on concerns about the government intruding into

the function of newspaper editors—concerns that “obviously are not present here.” Id. at 88. In Hurley, the Court

distinguished PruneYard by emphasizing that in PruneYard, it was unlikely that the messages would be attributed to

the owner, and that the owners could disavow those messages. Hurley v. Irish-American Gay, Lesbian and Bisexual

Group of Boston, 515 U.S. 557, 580 (1995).

361 PruneYard, 447 U.S. at 87.

362 547 U.S. 47, 63 (2006).

363 Id. at 51.

Congressional Research Service

35

Free Speech and the Regulation of Social Media Content

some law schools had restricted military recruiting on campus on the basis that the military,

through its “policy on homosexuals in the military,” violated the schools’ nondiscrimination

policies.364 FAIR argued that forcing the schools to “disseminate or accommodate a military

recruiter’s message” violated their First Amendment rights.365 The Court first noted that the

Solomon Amendment primarily regulated conduct and only incidentally compelled speech, in the

form of recruiting assistance such as sending emails or posting notices.366

Further, the Court held that “accommodating the military’s message does not affect the law

schools’ speech, because the schools are not speaking when they host interviews and recruiting

receptions.”367 Distinguishing Hurley, the Court said that “[u]nlike a parade organizer’s choice of

parade contingents, a law school’s decision to allow recruiters on campus is not inherently

expressive.”368 The Court said that “the expressive component” of the schools’ decisions to bar

military recruiters was “not created by the conduct itself but by the speech that accompanies

it.”369 Instead, as in PruneYard, the Court said that “[n]othing about recruiting suggests that law

schools agree with any speech by recruiters,” noting that the schools remained free to state that

they disagreed with the military’s policies.370

A number of federal trial courts have applied Tornillo to hold that search engines exercise

editorial judgment protected by the First Amendment when they make decisions about whether

and how to present specific websites or advertisements in search results.371 For example, in Zhang

v. Baidu.com, Inc., the trial court noted

This text is long and has been trimmed here. Open the source document for the complete record.