Content Moderation Regulation as Legal Role-Scripting

Indiana Law Journal Indiana Law Journal

Volume 99 Issue 4 Article 2

2024

Content Moderation Regulation as Legal Role-Scripting Content Moderation Regulation as Legal Role-Scripting

Sari Mazzurco

SMU Dedman School of Law

, [email protected]

Follow this and additional works at: https://www.repository.law.indiana.edu/ilj

Part of the Internet Law Commons

Recommended Citation Recommended Citation

Mazzurco, Sari (2024) "Content Moderation Regulation as Legal Role-Scripting,"

Indiana Law Journal

: Vol.

99: Iss. 4, Article 2.

Available at: https://www.repository.law.indiana.edu/ilj/vol99/iss4/2

This Article is brought to you for free and open access by

the Maurer Law Journals at Digital Repository @ Maurer

Law. It has been accepted for inclusion in Indiana Law

Journal by an authorized editor of Digital Repository @

Maurer Law. For more information, please contact

[email protected].

Content Moderation Regulation as Legal Role-Scripting

SARI MAZZURCO*

Lawmakers and scholars concerned with content moderation regulation typically

appeal to “analogies” to justify or undermine different forms of regulation. The logic

goes: law should afford individuals due process rights against speech platforms

because speech platforms are “like” speech governors as a matter of objective

reality. Other common analogies include common carriers, publishers, distributors,

shopping malls, and bookstores.

Commentators attempt to invoke social roles to understand what the content

moderation relationship is, what behaviors are “right” and “wrong” within it, and

how law should police behavioral deviations. But they do so without relying on

foundational sociology theory that explains what social roles are, what they do, and

how they come to be. Without this theoretical foundation, the discourse incompletely

portrays the project of content moderation regulation. Content moderation

regulations do not simply “take” speech platforms’ role as it currently exists––they

will also “make” speech platforms’ role by expressing that speech platforms should

be speech governors, common carriers, publishers, or something else, based on how

lawmakers choose to regulate.

This Article is the first to introduce role theory into the content moderation

discourse. Content moderation regulations are poised to define the basic contours

of what it means to be a “speech platform” because the role remains unsettled.

Earlier, the Communications Decency Act failed to articulate coherent roles within

the content moderation relationship. But current content moderation regulatory

reforms—including the PACT Act in Congress as well as state platform-common

carriage laws and their judicial review—have a renewed opportunity to script social

roles for speech platforms and individuals. Foregrounding these reforms’ role

scripts directs attention to urgent questions about whether they are likely to produce

a desirable content moderation relationship and an online speech ecosystem that

meets the public’s needs.

* Assistant Professor of Law, SMU Dedman School of Law. The author thanks for their

generous commentary Robert Post, Jack Balkin, Amy Kapczynski, Mark Lemley, Al

Klevorick, Vincent Mazzurco, Judge Thomas Ambro, and Spencer Livingstone. This Article

has benefitted from helpful feedback from participants in Cardozo Law School’s Intellectual

Property Scholars Conference, the Media Law & Policy Scholars Conference at The

University of Texas at Austin School of Law, the Works in Progress for Intellectual Property

Scholars Colloquium hosted by the University of Missouri School of Law and Saint Louis

University School of Law, the University of Michigan Law School Junior Scholars

Conference, the Yale Law School Freedom of Expression Scholars Conference, and the Yale

Information Society Project Writing Workshop. The views expressed in this Article are the

author’s own.

392239-ILJ 99-4_Text.indd 81392239-ILJ 99-4_Text.indd 81 5/29/24 10:48 AM5/29/24 10:48 AM

1132 INDIANA LAW JOURNAL [Vol. 99:1131

INTRODUCTION ..................................................................................................... 1132

I. SOCIAL ROLE AND CONTENT MODERATION REGULATION: THE PRESUMPTION OF

LINEARITY .................................................................................................... 1137

A. PRIMER ON CONTENT MODERATION ...................................................... 1137

B. SCHOLARLY DISCOURSE’S QUASI-ROLE LENS ....................................... 1138

C. THE MUTUAL CONSTITUTION OF LAW AND SOCIAL ROLES ................... 1141

II. TURNING THE LINE INTO A LOOP: LAW AND SOCIAL ROLES ON THE PLATFORM

INTERNET ...................................................................................................... 1142

A. PRIMER ON SOCIAL ROLES ..................................................................... 1143

B. LAW’S ROLE-SCRIPTING FUNCTION ....................................................... 1145

III. CDA SECTION 230 AS A LEGAL ROLE-SCRIPTING FAILURE .......................... 1150

A. SECTION 230 AS WRITTEN AND INTENDED ............................................ 1151

B. WHOM DOES SECTION 230 REGULATE AND WHOM DOES IT SERVE? .... 1152

C. ROLE EQUIVOCATION, ROLE CONFUSION, AND CONTENT MODERATION AS

“BUSINESS” PRACTICE ......................................................................... 1155

D. GONZALEZ V. GOOGLE’S RESIDUAL UNCERTAINTY ON SPEECH PLATFORMS

PUBLISHERS .................................................................................... 1159

IV. CONTENT MODERATION REGULATIONS’ ROLE SCRIPTS ................................ 1162

A. THE PACT ACT & THE PROTECT SPEECH ACT: SPEECH PLATFORMS AS

SPEECH GOVERNORS ........................................................................... 1163

B. TEXAS & FLORIDA PLATFORM COMMON CARRIAGE LAWS: SPEECH

PLATFORMS AS COMMON CARRIERS (MAYBE?) .................................. 1167

C. THE NETCHOICE CASES: SPEECH PLATFORMS AS CABLE OPERATORS .. 1172

CONCLUSION ........................................................................................................ 1175

NTRODUCTION

Months before Elon Musk purchased Twitter, he made his intentions clear. He

declared that he was a “free-speech absolutist” who would “set Twitter ‘free.’”

regarded Twitter as “the digital town square where matters vital to the future of

humanity are debated” and stated he was “against censorship that goes far beyond

the law.”

But beginning shortly after the deal closed, Musk began making Twitter

content moderation decisions that went far beyond the law. He suspended Kanye

West’s account after West posted an image of a swastika combined with a Star of

David;

he suspended a number of journalists from the Washington Post, New York

1. Khalid Albaih, Elon Musk, the Social Media Autocrat, AL JAZEERA (Jan. 1, 2023),

https://www.aljazeera.com/opinions/2023/1/1/2022-gave-us-a-new-social-media-autocrat

[https://perma.cc/TE9S-4ERC].

2. Maria Pasquini, Twitter Faces User Exodus After Elon Musk Deal—But Some High-

Profile Accounts Are Rapidly Gaining Followers, P

EOPLE (Apr. 27, 2022, 2:43 PM),

https://people.com/human-interest/twitter-faces-user-exodus-after-elon-musk-deal-but-some-

accounts-are-gaining-followers/ [https://perma.cc/XQG3-62RT].

3. Rachel Lerman, Cat Zakrzewski & Ellen Francis, Elon Musk Says Kanye West

Suspended from Twitter After Swastika Tweet, W

ASH. POST (Dec. 2, 2022, 1:19 PM),

https://www.washingtonpost.com/technology/2022/12/02/kanye-west-twitter-suspended-

elon-musk/ [https://perma.cc/FX4D-LYGW].

392239-ILJ 99-4_Text.indd 82392239-ILJ 99-4_Text.indd 82 5/29/24 10:48 AM5/29/24 10:48 AM

2024] CONTENT MODERATION REGULATION 1133

Times, and CNN for disclosing the location of Musk’s private jet (which is otherwise

publicly available information);

and he announced he would take down accounts

“created solely” to promote competing social networks.

More recently, he stated,

“New Twitter policy is to follow the science, which necessarily includes reasoned

questioning of the science.”

At many of these junctures, Musk took informal “polls”

of his followers, asking whether he should reinstate banned accounts or make other

changes.

On December 18, 2022, Musk again polled users, this time to ask whether

he should step down as CEO, and pledged to abide by the outcome.

Fifty-seven

percent of votes cast favored Musk stepping down,

but days later Musk said, “No

one wants the job who can actually keep Twitter alive. There is no successor.”

From the time Musk took Twitter over, he seemed to regard his role as a sort of

speech governor: one who exerts unilateral control over the public discourse that

flows through the speech platform, one who would be as responsive to Twitter users’

will as he pleased.

Users and journalists also seemed to relate to Musk’s content

moderation as a form of governance. Since his takeover, Insider Intelligence

predicted an “exodus” from Twitter by 2024,

and MIT Technology Review

reported Twitter might have lost a million users and counting.

Newsweek puzzled

4. Kurt Wagner, Davey Alba & Vlad Savov, Twitter Suspends Journalists Who Musk

Says Imperiled His Safety, B

LOOMBERG (Dec. 16, 2022, 12:30 AM),

https://www.bloomberg.com/news/articles/2022-12-16/twitter-suspends-accounts-of-

mastodon-journalists-covering-musk [https://perma.cc/P6P4-YN9E].

5. Susanne Barton, Twitter Will Remove Accounts That Link to Other Social Media,

LOOMBERG (Dec. 18, 2022, 12:56 PM), https://www.bloomberg.com/news/articles/2022-12-

18/twitter-will-remove-accounts-that-link-to-other-social-media [https://perma.cc/ZT8T-

WKT8].

6. Heather Hamilton, ‘Reasoned Questioning’ of Science: Musk’s New Twitter Policy

Met with Praise and Disdain, W

ASH. EXAM’R (Dec. 29, 2022, 10:41 AM),

https://www.washingtonexaminer.com/news/reasoned-questioning-science-elon-musk-new-

twitter-policy [https://perma.cc/T27U-S9HA].

7. Lerman et al., supra note 3.

8. Chantal Da Silva, Twitter Users Vote for Elon Musk to Step Down as CEO in Poll He

Launched, NBC

NEWS (Dec. 19, 2022, 7:53 AM), https://www.nbcnews.com/tech/social-

media/twitter-users-vote-elon-musk-quit-ceo-poll-rcna62332 [https://perma.cc/LR7H-J7Q4].

9. Id.

10. Edward Ludlow, Elon Musk Says ‘No One Wants’ Top Twitter Job, But Some People

Raise Their Hands, B

LOOMBERG (Dec. 19, 2022, 9:57 AM),

https://www.bloomberg.com/news/articles/2022-12-19/musk-says-no-one-wants-twitter-top-

job-but-some-people-pipe-up [https://perma.cc/RR2U-NSMY].

11. See supra notes 1–10 and accompanying text.

12. See Rob Pegoraro, Analysts Predict ‘Exodus’ of Twitter Users by 2024, PC

MAG.

(Dec. 13, 2022), https://www.pcmag.com/news/analysts-predict-exodus-of-twitter-users-by-

2024 [https://perma.cc/7L3B-WUMX].

13. See Chris Stokel-Walker, Twitter May Have Lost More Than a Million Users Since

Elon Musk Took Over, MIT T

ECH. REV. (Nov. 3, 2022),

https://www.technologyreview.com/2022/11/03/1062752/twitter-may-have-lost-more-than-

a-million-users-since-elon-musk-took-over/ [https://perma.cc/2P4Z-6XD7] (“The firm Bot

Sentinel, which tracks inauthentic behavior on Twitter by analyzing more than 3.1 million

accounts and their activity daily, believes that around 877,000 accounts were deactivated and

a further 497,000 were suspended between October 27 and November 1.”).

392239-ILJ 99-4_Text.indd 83392239-ILJ 99-4_Text.indd 83 5/29/24 10:48 AM5/29/24 10:48 AM

1134 INDIANA LAW JOURNAL [Vol. 99:1131

over whether Musk was a “savior come to rescue free speech . . . or [an] authoritarian

figure . . . who shuts down the free speech of his critics.”

Market researchers

suggested the #TwitterMigration was an act of “protest” in response to Musk’s

content moderation decisions.

Twitter users are not alone in relating to content moderation as a form of

governance. Digital ethnographers studying Twitch (a speech platform concerned

primarily with video game commentary) and Reddit also document banned users’

experience as one of governance.

In 2009, Facebook conducted its own experiment

with public participation in platform rulemaking, inviting users to vote on certain

policy changes.

Ultimately, it declined to follow the vote’s outcome.

Though speech governance has become a centerpiece in discourse on content

moderation,

it is one of many narratives percolating among the public, the press,

legal scholars, and lawmakers. Others have examined whether speech platforms

might be like common carriers,

publishers,

distributors,

shopping malls,

and

other entities,

or whether they are simply businesses, “profit-driven entities that

14. Katherine Brodsky, Elon Musk Is the New Trump: You Either Love Him or Hate Him,

EWSWEEK (Dec. 27, 2022, 5:37 PM), https://www.newsweek.com/elon-musk-new-trump-

you-either-love-him-hate-him-opinion-1769808 [https://perma.cc/6ZTQ-EV36].

15. Tim Chambers, A Snapshot of the X (Ex-Twitter) Migration, at 1–2, 10–14, Dewey

Square Group (2023), http://www.deweysquare.com/wp-content/uploads/2023/10/DSG-

Snapshot-of-the-XTwitter-Migration-2023-Q3.pdf [https://perma.cc/M2TE-4XSC]

(describing the link between Musk’s tweet-based content decisions and “#twittermigration”

from Twitter to other platforms); see also Stokel-Walker, supra note 13.

16. See Hibby Thach, Samuel Mayworm, Daniel Delmonaco & Oliver Haimson,

(In)visible Moderation: A Digital Ethnography of Marginalized Users and Content

Moderation on Twitch and Reddit, N

EW MEDIA & SOC’Y 1, 16 (July 18, 2022),

https://doi.org/10.1177/14614448221109804 [https://perma.cc/JNE8-P2TL].

17. See Facebook Opens Governance of Service and Policy Process to Users, F

ACEBOOK

(Feb. 26, 2009), https://about.fb.com/news/2009/02/facebook-opens-governance-of-

service-and-policy-process-to-users [https://perma.cc/EV54-VXZQ]; see also Sari Mazzurco,

Democratizing Platform Privacy, 31 F

ORDHAM INTELL. PROP. MEDIA & ENT. L. J. 792, 819–

20 (2021).

18. See Mazzurco, supra note 17, at 820.

19. See Evelyn Douek, Content Moderation as Systems Thinking, 136 H

ARV. L. REV. 526,

535–39 (2022); Evelyn Douek, Governing Online Speech: From “Posts-As-Trumps” to

Proportionality and Probability, 121 C

OLUM. L. REV. 759, 769–76 (2021); Aileen Nielsen,

The Rights and Wrongs of Folk Beliefs About Speech: Implications for Content Moderation,

27 UCLA

J.L. & TECH. 118, 121–30 (2022); Hannah Bloch-Wehba, Global Platform

Governance: Private Power in the Shadow of the State, 72 SMU

L. REV. 27, 33–42 (2019).

20. See Adam Candeub, Reading Section 230 As Written, 1 J.

FREE SPEECH L. 139, 146–

47 (2021); Christopher S. Yoo, The First Amendment, Common Carriers, and Public

Accommodations: Net Neutrality, Digital Platforms, and Privacy, 1 J.

FREE SPEECH L. 463,

500–04 (2021); Eugene Volokh, Treating Social Media Platforms Like Common Carriers?, 1

FREE SPEECH L. 377, 390–94 (2021).

21. See Volokh, supra note 20, at 403–07.

22. See id.

23. See id. at 415.

24. See id. at 415–27; Olivier Sylvain, Intermediary Design Duties, 50

CONN. L. REV.

203, 213, 234–35 (2018) (examining a “Good Samaritan” analogy).

392239-ILJ 99-4_Text.indd 84392239-ILJ 99-4_Text.indd 84 5/29/24 10:48 AM5/29/24 10:48 AM

2024] CONTENT MODERATION REGULATION 1135

moderate because it is in their business interests.”

Commentators assert they are

drawing analogies to figure out how law should regulate speech platforms, based on

whether platforms satisfy criteria that make them sufficiently “like” these other

entities.

Without acknowledging as much, commentators are in fact invoking social roles

of governor, common carrier, and publisher to understand what the content

moderation relationship is, what behaviors are “right” and “wrong” within it, and

how law should police deviations from behavioral expectations. But they are doing

so without building on foundational sociology theory on what social roles are, the

function they serve, and how they come to be. This Article contends, without this

theoretical foundation, the current discourse incompletely represents the project of

content moderation regulation. Content moderation regulations express that speech

platforms should be speech governors, common carriers, publishers, or something

else, based on how lawmakers choose to regulate. In other words, content moderation

regulation scripts speech platforms’ social roles.

This Article builds this essential theoretical architecture by introducing sociology

literature on social roles and legal scholarship on law and roles to the content

moderation discourse. Social roles typically help people navigate uncertain

interactions. Individuals’ roles, like priest or superintendent, and organizations’

roles, like school and military, make interactions meaningful by communicating

societal expectations about participants’ appropriate behaviors, interests,

characteristics, and values.

In emerging relationships—such as the content

moderation relationship between speech platforms and the public—roles can be

uncertain, causing confusion about how participants should treat one another.

decry that Twitter banned certain journalists (per Elon Musk’s orders) implies it

moderated content inappropriately. But whether it acted inappropriately depends

largely on how one understands Twitter’s social role, whether it’s a business, speech

governor, common carrier, or something else. Current regulatory reforms will supply

crucial content here. Law will help define what it means to be a “good” speech

platform through the rights, duties, behavioral constraints, and entitlements it assigns

to those who meet particular attributes.

Content moderation regulation is at a critical juncture. Texas and Florida

promulgated laws purporting to regulate speech platforms as common carriers.

The

Fifth Circuit and Eleventh Circuit have reached competing holdings about the

constitutionality of regulating speech platforms as common carriers, based on

divergent perspectives on whether speech platforms exercise “editorial judgment”

the First Amendment protects.

Congress is also considering whether to hold speech

25. Douek, Content Moderation as Systems of Thinking, supra note 19, at 559 (emphasis

omitted).

26. R

ALF DAHRENDORF, ESSAYS IN THE THEORY OF SOCIETY 36 (1968); Neal Gross et al.,

XPLORATIONS IN ROLE ANALYSIS 63 (1958).

27. See infra Part II.A.

28. H.B. 20, 87th Leg., 1st Spec. Sess. (Tex. 2021); S.B. 7072, 2021 Leg., Reg. Sess. (Fla.

2021).

29. See NetChoice, LLC v. Paxton, 49 F.4th 439, 445 (5th Cir. 2022) (reasoning “editorial

discretion” is not itself First Amendment-protected speech and disregarding the platforms’

claims they engage in protected “editorial discretion”); NetChoice, LLC v. Att’y Gen. Fla., 34

392239-ILJ 99-4_Text.indd 85392239-ILJ 99-4_Text.indd 85 5/29/24 10:48 AM5/29/24 10:48 AM

1136 INDIANA LAW JOURNAL [Vol. 99:1131

platforms accountable as governors when they take down user content.

That is to

say, at least five distinct legal authorities are poised to define speech platforms’

social role—potentially in tension with one another.

This Article draws insights from social role theory to help characterize the current

disordered socio-legal landscape and illuminate some repercussions likely to flow

from lawmakers’ regulatory decisions. It identifies Section 230 of the

Communications Decency Act (CDA)

as a legal role-scripting failure that led

courts to immunize an ever-widening scope of platforms’ conduct and left scholars

to puzzle over what exactly speech platforms are for the purpose of law. The sheer

number of social roles courts and scholars invoke to define “good” and “bad” content

moderation exemplifies the CDA’s failure to define the role-relationship.

Current content moderation regulatory reforms are another exercise in legal role-

scripting. Lawmakers’ choices to regulate speech platforms as speech governors,

common carriers, or publishers will guide platforms’ and individuals’ behavior,

address or neglect social harm, and support or undermine possible reforms in quite

different ways. For instance, the Eleventh Circuit’s characterization of speech

platforms like “cable operators” suggests that when Musk banned journalists from

Twitter, he did nothing wrong.

It is reasonable for publishers to make unilateral

decisions about what content they will and won’t publish. Directing attention to these

regulations’ role constructions channels debate to urgent questions about whether

they establish content moderation relationships and an online speech ecosystem that

meet the public’s needs.

This Article proceeds in four parts. Part I delves into the discourse on content

moderation, placing special focus on how legal scholars and courts commonly draw

role-based analogies to justify or undermine different forms of content moderation

regulation. Part II introduces sociological theory on social roles and presents law’s

role-scripting function. Part III relies on the theoretical architecture Part II builds to

identify Section 230 of the CDA as a failure of legal role-scripting that led to current

uncertainty about what role speech platforms play when they moderate content. Part

IV presents two current content moderation regulatory reforms as projects of legal

role-scripting. The Platform Accountability and Consumer Transparency (PACT)

Act and Protect Speech Act in Congress and Texas and Florida common carriage

laws (as well as courts reviewing those state laws) each construct different role-

relationships for platforms and people. Part IV enriches discourse on content

moderation regulation by examining the social meaning and legal pathways these

role constructions sustain.

F.4th 1196, 1209–10 (11th Cir. 2022) (holding platforms’ content moderation is “editorial

discretion” protected under the First Amendment).

30. See Protect Speech Act, H.R. 3827, 117th Cong. (2021); PACT Act, S. 4066, 116th

Cong. (2020).

31. 47 U.S.C. § 230 (1996).

32. NetChoice v. Att’y Gen. Fla., 34 F.4th at 1203.

392239-ILJ 99-4_Text.indd 86392239-ILJ 99-4_Text.indd 86 5/29/24 10:48 AM5/29/24 10:48 AM

2024] CONTENT MODERATION REGULATION 1137

I. SOCIAL ROLE AND CONTENT MODERATION REGULATION: THE PRESUMPTION OF

LINEARITY

In the current discourse on content moderation regulation, scholars and courts ask

what speech platforms are “like” when they moderate content: Speech governors?

Common carriers? Publishers? Or something else entirely? They attempt to fit

speech platforms into these categories to justify or undermine the imposition of

different sorts of legal rules to regulate content moderation. They do so typically by

examining whether speech platforms meet a set of objective and fixed criteria that

make them sufficiently “like” one of these other entities. When speech platforms fail

to meet enough of these criteria, they are deemed too different to merit that particular

form of regulation.

This Part presents the current discourse on content moderation, with a particular

focus on this typical, but incomplete way scholars and courts have engaged with

social roles. Commentators suggest the relationship between roles and law is linear,

with speech platforms’ role (as it exists in a sort of empirical reality) either

supporting or undermining the application of a particular form of law. But this is

only one side of the story: roles and law are in a feedback relationship, with law also

shaping what it means to be a speech platform. This Article asserts that side of the

feedback relationship is the core project of content moderation regulation—legal

role-scripting—though it has, so far, slipped under the radar.

A. Primer on Content Moderation

Professor Evelyn Douek gives a comprehensive account of content moderation.

At its most succinct, content moderation is “platforms’ systems and rules that

determine how they treat user-generated content on their services.”

Just below the

surface lies a web of complexity. She explains content moderation involves far more

than speech platforms’ decisions to take down particular pieces of content users

already posted. Content moderation also includes:

[I]ncreased reliance on automated moderation; sticking labels on posts;

partnerships with fact-checkers; . . . adding friction to how users share

content; giving users affordances to control their own online experience;

looking . . . to how users behave online to determine what should be

removed; and tinkering with the underlying dynamics of the very

platforms themselves.

Professor Olivier Sylvain explains platforms “design [their] application[s] to

elicit or shape user content or, conversely, ensure that certain kinds of content never

see the light of day.”

Professor Danielle Citron and Professor Neil Richards add

that often, platforms “tailor people’s online experiences,” including the content they

see, “based on fine-grained surveillance” about them.

An expansive range of

33. Douek, Content Moderation as Systems Thinking, supra note 19, at 528 n.2.

34. Id. at 531 (emphasis omitted).

35. Sylvain, supra note 24, at 224.

36. Danielle Keats Citron & Neil M. Richards, Four Principles for Digital Expression

392239-ILJ 99-4_Text.indd 87392239-ILJ 99-4_Text.indd 87 5/29/24 10:48 AM5/29/24 10:48 AM

1138 INDIANA LAW JOURNAL [Vol. 99:1131

speech platforms engage in content moderation. Beyond Facebook, Google, and

Twitter, platforms like Tinder,

Reddit,

Wikipedia,

Dropbox,

Amazon,

Yelp,

and even the New York Times

and the Washington Post

in their comments

sections engage in content moderation.

Citron and Richards explain, on the whole, content moderation entails “[p]rivate

power over digital expression” that “should be paired with responsibility to the

public.”

But what, exactly, should those responsibilities be? As it turns out, that

question is at the center of debates on content moderation regulation, with

commentators typically justifying or undermining particular responsibilities based

on whether speech platforms are sufficiently “like” certain other entities.

B. Scholarly Discourse’s Quasi-Role Lens

As in the “Muskification” of Twitter, speech governance has become a dominant

narrative in discourse on content moderation regulation. Professor Hannah Bloch-

Wehba puts the logical connection succinctly: “When Internet platforms engage in

policing and moderating content online, they engage in a form of private

governance,” albeit one that “suffer[s] from a ‘democratic deficit.’”

“Like

administrative agencies, platforms set rules by promulgating internal regulations

about user speech. . . . [A]s the number of disputes . . . grows . . . platforms

increasingly create adjudicatory and appellate mechanisms,”

like Facebook’s

(You Won’t Believe #3!), 95 WASH. U. L. REV. 1353, 1356 (2018).

37. Community Guidelines, T

INDER, https://policies.tinder.com/community-

guidelines/intl/en/ [https://perma.cc/QR3X-8EYD].

38. Reddit Content Policy,

REDDIT, https://www.redditinc.com/policies/content-policy

[https://perma.cc/TQ2T-VUPQ].

39. Wikipedia: Automated Moderation, W

IKIPEDIA,

https://en.wikipedia.org/wiki/Wikipedia:Automated_moderation [https://perma.cc/K9UL-

DVHY].

40. Dropbox Acceptable Use Policy, D

ROPBOX,

https://www.dropbox.com/acceptable_use [https://perma.cc/5GB8-L4PH].

41. Community Guidelines, A

MAZON.COM,

https://www.amazon.com/gp/help/customer/display.html?nodeId=GLHXEX85MENUE4XF

[https://perma.cc/HE97-S2CL].

42. Content Guidelines, Y

ELP, https://www.yelp.com/guidelines

[https://perma.cc/6LXW-6VXP].

43. The Comments Section, N.Y.

TIMES, https://help.nytimes.com/hc/en-

us/articles/115014792387-The-Comments-Section [https://perma.cc/39VS-3ZQ6].

44. Discussion and Submission Guidelines, W

ASH. POST (Nov. 23, 2021, 12:00 PM),

https://www.washingtonpost.com/discussions/2021/11/23/discussion-submission-guidelines/

[https://perma.cc/WTD4-U6EN].

45. See Enrique Armijo, Reasonableness as Censorship: Section 230 Reform, Content

Moderation, and The First Amendment, 73

FLA. L. REV. 1199, 1212 (2021) (explaining that

numerous entities not commonly regarded as platforms also host user content and engage in

content moderation to varying degrees).

46. Citron et al., supra note 36, at 1383.

47. Bloch-Wehba, supra note 19, at 33.

48. Id. at 29–30.

392239-ILJ 99-4_Text.indd 88392239-ILJ 99-4_Text.indd 88 5/29/24 10:48 AM5/29/24 10:48 AM

2024] CONTENT MODERATION REGULATION 1139

Oversight Board.

Law should then “ensur[e] democratic accountability and

legitimacy,” through “the application of administrative law principles and values to

hold platforms to account.”

A number of other scholars draw similar analogies to

justify the promulgation of public-law-type content moderation regulations. Douek

similarly makes a connection to administrative law.

Professor Rory Van Loo

suggests a likeness to the judiciary, calling for legally mandated platform

procedure.

On the other side of the coin, numerous scholars (including some in favor of these

reforms) point out where these analogies break down. Professor Kyle Langvardt

argues the governance analogy is inappropriate because speech platforms aren’t

chosen through democratic processes.

Douek raises that the scale and speed of

content moderation make the implementation of additional individual procedural

rights infeasible.

Douek also asserts that “[p]latforms are businesses [that] will

prioritize their bottom line”

and content moderation is the “commodity [they]

offer” because it is in their “business interest.”

Douek seems to suggest platforms

are not governors and should not be required to satisfy individual rights because they

are actually, as a matter of empirical reality, businesses. If content moderation

regulation must take speech platforms as businesses, any perception of a democratic

deficit is mistaken and an inappropriate driver for law. By extension, both individual

rights and administrative law remedies would be misplaced.

The next most prominent narrative is common carriage. Professor Christopher

Yoo provides the most comprehensive account of what it means to be a common

carrier and whether speech platforms meet those criteria.

He examines Justice

Thomas’s opinion in Biden v. Knight First Amendment Institute,

which represents

“the types of considerations that have historically been used to define common

carriers.”

These include: [M]arket power, whether an industry is “affected with the

public interest,” whether the entity regulated is part of the transportation or

communications industry, whether it receives countervailing benefits from the

government, and whether the actor holds itself out as providing service to all.

concludes that each of these criteria has been applied inconsistently, fails to

determine an entity’s status as a common carrier, or is easily evaded. Ultimately, he

49. See Rory Van Loo, Federal Rules of Platform Procedure, 88 U. CHI. L. REV. 829, 870

(2021) (recommending that Congress require large platforms to have external oversight

boards, like Facebook’s Oversight Board).

50. Bloch-Wehba, supra note 19, at 33.

51. See Douek, Content Moderation as Systems Thinking, supra note 19, at 532–33.

52. See Van Loo, supra note 49, at 833.

53. See Kyle Langvardt, Regulating Online Content Moderation, 106 G

EO. L.J. 1353,

1358 (2018).

54. See Douek, Content Moderation as Systems Thinking, supra note 19, at 532.

55. Id. at 589.

56. Douek, Governing Online Speech: From “Posts-As-Trumps” to Proportionality and

Probability, supra note 19, at 768, 806.

57. See generally Yoo, supra note 20.

58. Id. at 465 (citing 141 S. Ct. 1220, 1222–23 (2021) (Thomas, J., concurring)).

59. Id.

60. Id. at 465–66.

392239-ILJ 99-4_Text.indd 89392239-ILJ 99-4_Text.indd 89 5/29/24 10:48 AM5/29/24 10:48 AM

1140 INDIANA LAW JOURNAL [Vol. 99:1131

finds speech platforms engage in too much editorial discretion when it comes to

content moderation, such that they don’t hold themselves out as serving all members

of the public.

Professor Eugene Volokh, on the other hand, proposes platforms

might be close enough to common carriers when they host user content.

He argues

advocacy groups depend on speech platforms to communicate with the public, and

“denying a group a vastly important means of public communication is a serious

burden.”

He also suggests, much like e-mail services, users neither want nor expect

platforms to screen what a user posts to their own page based on the viewpoint

expressed.

Part IV.B discusses, in greater detail, the Eleventh Circuit and Fifth Circuit

decisions on Florida and Texas’s platform common carriage laws. The courts’

opinions also wrestle with whether platforms are sufficiently “like” common carriers

to justify their regulation as such. The Eleventh Circuit found platforms are not like

common carriers because they aren’t “dumb pipes” that simply transmit data from

point A to point B.

Rather, it reasoned platforms exercise “editorial judgment”

protected by the First Amendment.

The Fifth Circuit, by contrast, found speech

platforms’ content moderation practices implicate the public interest.

It reasoned

that the platforms host public discourse and they hold themselves out to serve the

public by “permit[ting] any adult to make an account and transmit expression after

agreeing to the same boilerplate terms of service.”

The third most prominent narrative in content moderation discourse concerns

whether speech platforms engage in “editorial discretion” (also called “editorial

judgment”)

such that content moderation is itself a form of expression the First

Amendment protects from government regulation. Rather than focus on a single

analogy like democratic government or common carriage, scholars and courts

instead assess whether speech platforms exercise editorial discretion by analogizing

to and distinguishing from a number of differently regulated entities, including

newspapers, broadcasters, editors, publishers, book stores, law schools, malls, and

parades.

This is, in a sense, an outgrowth from the case method all law students

learn in their first year. Legal reasoning and argument typically rely on compelling

analogies or disanalogies to prior precedent.

What’s notable, here, is scholars’ and courts’ analogies center on speech

platforms’ likeness to a particular set of entities, as opposed to other facts. For

example, Professor Volokh engages in a rigorous analysis of whether platforms are

sufficiently “like” any of these entities to warrant similar First Amendment treatment

by examining whether they share distinctive features.

He concludes, for instance,

61. See id. at 465–75.

62. See Volokh, supra note 20, at 408.

63. Id. at 390.

64. See id. at 386.

65. NetChoice, LLC v. Att’y Gen., Fla., 34 F.4th 1196, 1204 (11th Cir. 2022).

66. See id. at 1209.

67. NetChoice, LLC v. Paxton, 49 F.4th 439, 473 (5th Cir. 2022).

68. Id. at 474.

69. See NetChoice v. Att’y Gen., Fla., 34 F.4th at 1203.

70. See supra notes 19–25 and accompanying text.

71. See Volokh, supra note 20, at 403, 415–24.

392239-ILJ 99-4_Text.indd 90392239-ILJ 99-4_Text.indd 90 5/29/24 10:48 AM5/29/24 10:48 AM

2024] CONTENT MODERATION REGULATION 1141

that platforms’ hosting function makes them unlike newspapers or broadcasters

because: (1) platforms don’t have the same time or space constraints when it comes

to the content they make available; (2) users don’t rely on platforms (as they do

newspapers and broadcasters) to exclude material they find offensive or useless; and

(3) platforms aren’t a coherent and consistent speech product.

Using a similar form

of reasoning, the Eleventh Circuit found:

Just as the parade organizer exercises editorial judgment when it refuses

to include in its lineup groups with whose messages it disagrees, and just

as a cable operator might refuse to carry a channel that produces content

it prefers not to disseminate, social-media platforms regularly make

choices ‘not to propound a particular point of view.’

In effect, the court held platforms merit the same First Amendment protection as

parade organizers and cable operators because they are sufficiently “like” those

entities.

C. The Mutual Constitution of Law and Social Roles

Bloch-Wehba, arguing content moderation regulation should draw from

administrative law, asserts “rather than debating about the nature of platforms’

power, the right question is whether the exercise of that power is legitimate . . . .”

But, whether the power platforms exert when they moderate content is legitimate

depends quite a lot on the role one perceives the platform is playing. The scholarly

discourse suggests that platforms’ role is unsettled, both as a matter of social

understanding and for the purposes of law.

This Part presents the scholarly debate on content moderation regulation to

highlight a key, common oversight: commentators in the academy and the courts

presume the relationship between speech platforms’ social role and law is static and

linear. The argument goes, speech platforms are like X, and so law is justified

regulating them as though they are X. (Or, inversely, speech platforms are not like

X, and so law isn’t justified regulating them like X.)

Much effort is placed on determining whether speech platforms are like governors,

common carriers, newspapers, etc. by connecting the “empirical realities”

of how

platforms behave to objective, fixed criteria that make any entity a governor,

common carrier, newspaper, etc.

72. See id. at 404–05.

73. NetChoice v. Att’y Gen., Fla., 34 F.4th at 1213 (quoting Hurley v. Irish-American

Gay, Lesbian and Bisexual Group of Boston, 515 U.S. 557, 575 (1995)).

74. Bloch-Wehba, supra note 19, at 68.

75. Douek, Governing Online Speech: From “Posts-As-Trumps” to Proportionality and

Probability, supra note 19, at 767.

Social roles Law

392239-ILJ 99-4_Text.indd 91392239-ILJ 99-4_Text.indd 91 5/29/24 10:48 AM5/29/24 10:48 AM

1142 INDIANA LAW JOURNAL [Vol. 99:1131

Commentators are attempting to approach content moderation regulation from a

social role lens, but without relying on a theoretical architecture that lays out on what

roles are, the function they serve in social life, and their mutually constitutive

relationship with law.

As the next Part explains in greater detail, social roles are

not lists of essential physical, technical, or economic criteria law responds to when

making regulatory decisions. Rather, roles are a shorthand for the expectations

society holds for entities’ appropriate behavior, values, interests, and attributes in

various contexts. Roles are constantly in flux, open to contest, and partly constituted

by law. That is to say, law will influence what it means to be a speech platform based

on how lawmakers decide to regulate content moderation, just as law continually

shapes what it means to be a governor, common carrier, newspaper, etc. by how it

characterizes and regulates those roles.

This theoretical reorientation suggests that, rather than focus on crafting the perfect

analogy for law as a matter of observing objective reality, scholars and lawmakers

ought to place more attention on whether content moderation regulation sets a

desirable expectation of what speech platforms should be like.

II.

TURNING THE LINE INTO A LOOP:

AW AND SOCIAL ROLES ON THE PLATFORM INTERNET

This Part constructs the theoretical architecture on which the remainder of the

Article rests. It draws insights from sociology theory on social roles—what social

roles are, the function they serve, and how they come to be—to fill a gap in the

current discourse on content moderation regulation. Social roles supply societal

expectations about an entity’s appropriate attributes, interests, values, and behaviors

76. C.f. Sylvain, supra note 24, at 277. Professor Olivier Sylvain provides an exceptional

account of how the CDA failed to deliver on its ostensible aim to motivate interactive

computer services to act as Good Samaritans—and receive concomitant immunity from

liability for any harm done in the course of rendering aid. Id. at 213. Professor Sylvain

recommends that amending Section 230(c)(1)’s liability shield to cover only providers

“voluntar[y] act[ions] in good faith as Good Samaritans” might motivate speech platforms to

act as Good Samaritans. Id. at 277, 214. Without drawing an overt connection to role theory,

Professor Sylvain recognizes that law’s treatment of its covered entity (i.e., interactive

computer services) has the capacity to shape speech platforms’ role, as a Good Samaritan or

something else entirely.

Social roles

Law

392239-ILJ 99-4_Text.indd 92392239-ILJ 99-4_Text.indd 92 5/29/24 10:48 AM5/29/24 10:48 AM

2024] CONTENT MODERATION REGULATION 1143

in a particular relationship context. Law often contributes to these expectations based

on how it defines roles and regulates entities that meet definitional parameters. Law

is especially influential when social roles are otherwise unsettled or uncertain, as is

the case for speech platforms. Placing focus on this side of the feedback loop—law’s

role-scripting function—suggests content moderation regulations do not simply

“take” speech platforms’ role (and regulate on that basis) but they “make” speech

platforms’ role based on how they regulate content moderation.

A. Primer on Social Roles

As individuals navigate their daily lives, they interact with others who occupy any

number of social roles: for instance, the teachers who educate their children, the

landlords who collect their rent, or the baristas who make their lattes. For

sociologists, social roles are a “building block” of social reality because they make

everyday social interactions meaningful and relatively predictable.

Social roles

serve as a kind of shorthand for a set of expectations society holds for actors’

appropriate behavior, values, interests, and attributes in various contexts.

Though the concept of social roles may seem abstract, in practice, social roles are

the lenses through which individuals see the world.

One might expect teachers to

help students learn different subjects, but not give them medical advice. It’s

reasonable to expect teachers to have achieved a high level of education in the field

of education, and not be registered sex offenders. These expectations—also called

norms—help individuals evaluate the meaning of others’ actions based on whether

they conform or diverge.

If a teacher does not meet these expectations, it’s

reasonable to regard his conduct as “inappropriate” or consider him a “bad” teacher.

Roles also help individuals figure out how to treat others.

A teacher should know,

because of his social role and that of his students, that he should not give his students

medical advice.

77. See Meir Dan-Cohen, Between Selves and Collectivities: Toward a Jurisprudence of

Identity, 61 U.

CHI. L. REV. 1213, 1218–19, 1228–29 (1994); see also PETER L. BERGER &

THOMAS LUCKMANN, THE SOCIAL CONSTRUCTION OF REALITY 60–61, 74–76 (1967).

78. See Frank Dobbin, Economic Sociology, in T

WENTY-FIRST CENTURY SOCIOLOGY: A

REFERENCE HANDBOOK 320 (Clifton D. Bryant & Dennis L. Peck eds., 2007) (explaining

Émile Durkheim’s view that social role shapes economic behavior); D

AHRENDORF, supra note

26, at 35–37 (explaining social roles carry expected modes of behavior, characteristics, beliefs,

and interests); B.J. Biddle, Recent Development in Role Theory, 12 A

NN. REV. SOCIO. 67, 70–

71 (1986) (cataloguing sociology scholarship on functional and symbolic interactionist role

theory); G

ROSS ET AL., supra note 26, at 59–60, 63 (explaining that a role is a set of

expectations about someone’s attributes and behaviors); John Scott, Status and Role:

Structural Aspects, in I

NT’L ENCYCLOPEDIA SOC. & BEHAV. SCIS. (Neil Smelser & Paul Balies

eds., 2001) (synthesizing sociology literature on role-taking and role-making).

79. See G

EORGE A. AKERLOF & RACHEL E. KRANTON, IDENTITY ECONOMICS: HOW OUR

IDENTITIES SHAPE OUR WORK, WAGES, AND WELL-BEING 11 (2010).

80. See Peter M. Hall, A Symbolic Interactionist Analysis of Politics, 42 S

OCIO. INQUIRY

35, 38–40 (1972); D

AHRENDORF, supra note 26, at 44; Ralph H. Turner, Role Theory, in

ANDBOOK OF SOCIOLOGICAL THEORY 233, 235 (Jonathan H. Turner ed., 2001).

81. See Hall, supra note 80, at 39–40, 55; Turner, supra note 80, at 235.

392239-ILJ 99-4_Text.indd 93392239-ILJ 99-4_Text.indd 93 5/29/24 10:48 AM5/29/24 10:48 AM

1144 INDIANA LAW JOURNAL [Vol. 99:1131

Organizations of people can also have social roles on an organization-wide level.

For instance, one can talk about the ways it’s appropriate for a school to discipline a

student, regulate student or teacher speech, or facilitate religious practice.

It would

be reasonable to expect a school, the military, and a church to regulate speech quite

differently, on the basis of their different social roles.

Social roles typically arise through a process of continuous interaction in

society—between and among individuals, organizations, governments, and others.

The process is dialectical, meaning it involves persistent conflict between opposing

ideas about what social roles are and the expectations we should associate with

them.

Still, in normal times, social roles and their associated norms are well known

and form the assumed, background rules against which individuals operate.

Individuals can understand what it means to be a school before ever interacting with

one because they have access to cultural knowledge about a school’s appropriate

attributes and behaviors.

People learn social roles in society through the process of socialization; that is,

by learning their culture from the time they are born.

Socialization also helps

perpetuate social roles. When someone internalizes a social role—when they accept

it as frame for their own behavior—they are likely to conform to it and carry it

forward in their culture.

Viviana Zelizer gives the example of payments for

household chores within a family: if children accept that, by virtue of their role, they

are expected to participate in household chores without payment, they are likely to

comply and carry the same expectations for their future children.

Beyond

socialization, forms of sanctions, including social shaming and ostracism, rewards,

and legal penalties, help sustain social roles in society.

When social roles like “speech platform” are uncertain, individuals are limited in

their ability to understand what constitutes a “normal” relationship with a speech

platform. How is it appropriate for a speech platform to behave? What individual

interests are reasonable within this relationship? How should individuals tailor their

behavior to match the relationship’s context? For instance, when Elon Musk took

82. See, e.g., Kennedy v. Bremerton Sch. Dist., 597 U.S. 507, 525–26 (2022).

83. See J

EFFREY K. HASS, ECONOMIC SOCIOLOGY: AN INTRODUCTION 9 (2006); Mark

Granovetter, Economic Action and Social Structure: The Problem of Embeddedness, 91 A

J. SOCIO. 481, 486 (1985); NEIL FLIGSTEIN, THE ARCHITECTURE OF MARKETS 27–28 (2001).

84. See B

ERGER ET AL., supra note 77, at 61. Ideas about social roles may also differ

dramatically between different communities. One could imagine, for instance, LGBTQ ideas

of what it means to be a parent, father, or mother diverging from heteronormative ideas about

these roles.

85. See F

LIGSTEIN, supra note 83, at 27.

86. See Hall, supra note 80, at 37–38; D

AHRENDORF, supra note 26, at 26–27, 56; Scott,

supra note 78; B

ERGER ET AL., supra note 77 at 74.

87. See Hall, supra note 80, at 38; B

ERGER ET AL., supra note 77, at 74; DAHRENDORF,

supra note 26, at 56.

88. See Elizabeth Anderson, Beyond Homo Economicus: New Developments in Theories

of Social Norms, 29 P

HIL. & PUB. AFFS. 170, 191 (2000) (citing Viviana Zelizer, How Do We

Know Whether a Monetary Transaction is a Gift, an Entitlement, or Compensation?, in

CONOMICS, VALUES, AND ORGANIZATION 329–31 (Avner Ben-Ner & Louis Putterman eds.,

1998)).

89. See D

AHRENDORF, supra note 26, at 38, 42–43.

392239-ILJ 99-4_Text.indd 94392239-ILJ 99-4_Text.indd 94 5/29/24 10:48 AM5/29/24 10:48 AM

2024] CONTENT MODERATION REGULATION 1145

down Kanye West’s antisemitic Twitter post,

it’s unclear what the take down

meant, let alone whether it was legitimate. Was it a despotic act of censorship,

discrimination on the basis of West’s viewpoint, or an editorial judgment? Should

West petition the platform for redress, sue for discrimination, or accept the

platform’s judgment? Without an understanding of Twitter’s social role when it

moderates content, individuals lack the cultural knowledge that would enable them

to ascribe a particular meaning to Musk’s act and react accordingly.

B. Law’s Role-Scripting Function

Citron and Richards write, “How courts talk about the Internet matters, not only

in how they decide individual cases, but also in how they frame similar issues for

future courts.”

When it comes to content moderation regulation, the statements law

makes about speech platforms matter quite a lot, not only for lawmakers’ purposes,

but also to establish baseline expectations of what it means to be a speech platform

and how one behaves appropriately within that role.

Though this is a new lens to view the project of content moderation, it is a deep-

rooted function of law.

Law helps script social roles—it engages in “legal role-

scripting”—when it makes statements about the characteristics and appropriate

behaviors of the entities it regulates and the public it serves. Law is especially able

to exert influence over social roles that are unsettled or uncertain—like “speech

platform.”

Content moderation regulations’ role-scripting can provide individuals

the cultural knowledge to evaluate whether speech platforms are treating them the

way they should, to know how they should act toward speech platforms, and to

modify the relationship’s social obligations as it continues to unfold. Regulations’

role choices are also likely to set further lawmaking down particular paths and close

off others.

Legal role-scripting is both a key function and a common regulatory mechanism

of modern liberal democracies.

Some laws, such as much of criminal law, proscribe

90. Twitter, Instagram Block Kanye West over Antisemitic Posts, ASSOCIATED PRESS

(Oct. 10, 2022, 5:39 PM), https://apnews.com/article/twitter-inc-entertainment-music-

ba5c710ec59d195fe4d83cb2c9343589 [https://perma.cc/BV48-9PCH].

91. Citron et al., supra note 36, at 1383.

92. See Cass R. Sunstein, Social Norms and Social Roles, 96 C

OLUM. L. REV. 903, 921,

923 (1996) (describing the manner in which “law is often self-consciously concerned with

social roles”); Dan-Cohen, supra note 77, at 1228–29 (describing the ways in which law

“scripts” social roles); Eric J. Mitnick, Law, Cognition, and Identity, 67 L

A. L. REV. 823, 824,

831 (2007) (explaining how “legal rules . . . categorize persons”); Bert I. Huang, Law and

Moral Dilemmas, 130 H

ARV. L. REV. 659, 678 (2016) (“Another, more subtle way that the law

might influence moral intuitions is by defining official or social roles, which in turn set our

expectations about correct or blameworthy behavior.”).

93. The widespread use of the term “platform” to refer to entities that intermediate

information and relationships is itself significant. See Julie E. Cohen, Law for the Platform

Economy, 51 U.C.

DAVIS L. REV 133, 145 (2017) (“In Tarleton Gillespie’s formulation, the

term ‘platform’ appears to offer users a ‘raised, level surface’ on which to present themselves,

but at the same time it elides the necessary work of defining and policing the platform’s

edges.”) (internal citation omitted).

94. See Mitnick, supra note 92, at 824 (“[M]odern liberal democratic legal institutions

392239-ILJ 99-4_Text.indd 95392239-ILJ 99-4_Text.indd 95 5/29/24 10:49 AM5/29/24 10:49 AM

1146 INDIANA LAW JOURNAL [Vol. 99:1131

certain acts regardless of the actor’s social role. Whether someone is a cashier or an

uncle, they cannot intentionally kill another person (without justification).

But,

more often than not, the law treats entities categorically through its choice of legal

subject. The rights, responsibilities, behavioral constraints, and entitlements law

gives its legal subject form the basic outline of a social role.

Some scholars suggest that law can only—imperfectly—reflect settled social

roles,

but others acknowledge law historically has created a number of social roles

altogether or redefined substantially the behavioral norms associated with particular

roles.

Harlan Fiske Stone has noted, for example, family law’s extensive

elaboration of “husband” and “wife” roles through its allocations of rights and

responsibilities from the time of the nation’s founding.

Legislation gradually

ushered in wives’ legal independence and new, progressive norms within the

husband-wife relationship.

100

The origin of the corporate form and its jurisprudential development exemplify

the law’s ability to create social roles wholesale and continually reshape them.

101

When a government granted a corporate charter, the new entity that emerged—the

corporation—had a set of privileges and responsibilities granted by the state, not only

to those who occupied the new social role of “shareholder,” but to the public.

102

Corporate law both defined the social role of “corporation” and constructed a suite

persist unreservedly in structuring the categories through which individuals perceive social

life and status.”); Manfred Rehbinder, Status, Contract, and the Welfare State, 23 S

TAN. L.

REV. 941, 955 (1971) (describing modern law as “a law of roles preformed and safeguarded

by the state, yet open and subject to constant change”).

95. See, e.g., N.Y.

PENAL LAW § 125.27 (2024),

https://www.nysenate.gov/legislation/laws/PEN/125.27 [https://perma.cc/AML8-7BAY].

96. See Mitnick, supra note 92, at 824, 828, 831, 833, 868–69 (examining the

“constitutive nature of law” in terms of how it supplies social meaning to roles based on the

manner it regulates people occupying those roles); Rehbinder, supra note 94, at 951–52 (“The

kind of law that differentiates according to social position of the individual is a law that

involves social roles.”); Dan-Cohen, supra note 77, at 1228–30 (describing the numerous ways

law helps script social roles); Huang, supra note 92, at 678 (examining how legally afforded

rights and responsibilities influence people’s moral understanding of role-based social

responsibilities).

97. See H

ARLAN F. STONE, LAW AND ITS ADMINISTRATION 34 (1915) (suggesting the

social significance of legal changes to wives’ status); Paul Bohannan, The Differing Realms

of the Law, 67 A

M. ANTHROPOLOGIST 33, 35–37 (1965) (contending law is always out-of-

phase with society and a step behind role-based social norms).

98. See Sunstein, supra note 92, at 923; Dan-Cohen, supra note 77, at 1229–30; Mitnick,

supra note 92, at 824, 865; S

TONE, supra note 97, at 78–79, 82; Neil Fligstein, From the

Transformation of Corporate Control, in T

HE NEW ECONOMIC SOCIOLOGY: A READER 408–09

(Frank Dobbin ed., 2004) (describing corporations as legal constructs comprised of many, new

roles); William G. Roy, Socializing Capital: The Rise of the Large Industrial Corporation in

America, in T

HE NEW ECONOMIC SOCIOLOGY: A READER 438–39, 450–51 (Frank Dobbin ed.,

2004) (describing the historical elaboration of corporations’ social role).

99. S

TONE, supra note 97, at 78–79.

100. Id.

101. See id. at 82; Fligstein, supra note 98, 408–09; Roy, supra note 98, at 438–39, 450–

51.

102. See Fligstein, supra note 98, 408–09; Roy, supra note 98, at 450–51.

392239-ILJ 99-4_Text.indd 96392239-ILJ 99-4_Text.indd 96 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1147

of social roles within the corporation, each with a set of legally scripted behavioral

obligations: shareholder, officer, director, chair, etc.

103

Since then, the Supreme

Court has recognized corporations’ rights to speak, fund electioneering

communications, and practice religion, suggesting these are all normal social

behaviors for corporations.

104

Even when law supplies roles’ initial social meaning, those roles are continually

redefined as actors interact in the roles and with the roles. One need only look, for

example, at the syllabus of the Harvard Business School course “Corporate

Governance and Boards of Directors.” The legal rights and obligations that attach to

the many social roles within the corporate structure are only a small component of

the course. Rather, the course focuses on “the complex dynamics among boards,

executives, and shareholders,” and the “managerial[] and behavioral issues that

directors must contend with. . . .”

105

Society and law interact in a constant feedback loop. Contest over the meaning of

social roles and actual social behaviors continually reshape the expectations

associated with legally constructed social roles. When societal expectations

eventually diverge from law’s formulation of particular roles, social movements

often call for law to catch up. For instance, the Environmental, Social, and

Governance (ESG) movement is pressing lawmakers to impose a range of new

regulations on corporations—from increasing board diversity to reducing carbon

emissions—that reflect emerging societal expectations of how corporations should

behave.

106

Different actors in society may have different ideas of what roles are relevant and

desirable in any new relationship. The state of public discourse on content

moderation, and scholarly discourse on its regulation, reflect that notions of speech

platform’s role are unsettled, uncertain, and deeply contested. Survey evidence

reveals, for instance, that Democrats broadly approve speech platforms’ political

fact-checking, whereas Republicans broadly disapprove of the practice.

107

Whereas

some of the largest speech platforms have begun to represent themselves as

103. See, e.g., N.Y. BUS. CORP. LAW art. 6 (2014),

https://www.nysenate.gov/legislation/laws/BSC/A6 [https://perma.cc/27U2-GZ8Z]; id. at art.

7 (2017), https://www.nysenate.gov/legislation/laws/BSC/A7 [https://perma.cc/Q2BG-

RXA4].

104. See Citizens United v. FEC, 558 U.S. 310, 365 (2010); Burwell v. Hobby Lobby

Stores, Inc., 573 U.S. 682, 707 (2014).

105. See Corporate Governance and Boards of Directors, H

ARV. BUS. SCH.: COURSE

CATALOG, https://www.hbs.edu/coursecatalog/2010.html [https://perma.cc/R38Z-2GBQ].

106. See Leah Malone, Emily Holland & Carolyn Houston, ESG Battlegrounds: How the

States Are Shaping the Regulatory Landscape in the U.S., H

ARV. L. SCH. F. ON CORP.

GOVERNANCE (Mar. 11, 2023), https://corpgov.law.harvard.edu/2023/03/11/esg-

battlegrounds-how-the-states-are-shaping-the-regulatory-landscape-in-the-u-s/

[https://perma.cc/A476-XDDW].

107. Most Americans Think Social Media Sites Censor Political Viewpoints, P

EW RSCH.

CTR. (Aug. 19, 2020), https://www.pewresearch.org/internet/2020/08/19/most-americans-

think-social-media-sites-censor-political-viewpoints/ [https://perma.cc/9BKD-UMGF].

392239-ILJ 99-4_Text.indd 97392239-ILJ 99-4_Text.indd 97 5/29/24 10:49 AM5/29/24 10:49 AM

1148 INDIANA LAW JOURNAL [Vol. 99:1131

“governors,”

108

conservative politicians insist they are “common carriers,”

109

and

numerous legal scholars argue they are actually “publishers.”

110

Whether law creates social roles or crystallizes those already present in society,

law’s involvement in role-scripting carries a special force in society. First, the public

readily internalizes legally constructed roles and norms, such that individuals are

rarely conscious of the law’s influence on their perceptions.

111

For example, in an

experimental study examining reactions to different iterations of tort law’s classic

trolley problem, Bert Huang found that respondents’ views of various roles’ moral

responsibilities changed depending on the legal duty the law assigned to each role.

112

Individuals often instinctively translate legal definitions into social norms, especially

when norms are not settled.

113

Second, law has a special ability to enforce social roles

and norms through the threat of sanctions for noncompliance.

114

Feminist activists

who fought for Congress to pass the Civil Rights Act of 1991 sought to leverage the

force of legal sanctions to drive changes to workplace behavioral norms toward

women.

115

Content moderation regulation is poised to mediate among the many conceptions

of speech platforms’ role. It may provide the role’s preliminary contours through the

duties it imposes, the entitlements it affords, and the statements it makes about the

legal attributes of speech platforms’ content moderation relationship with

individuals. The unsettled status of speech platforms’ social role suggests law may

be especially powerful in institutionalizing those social roles. Society may more

readily take up legal definitions in the absence of well-established alternatives. The

potential for legal sanctions is also crucial in this context. For instance, Google would

108. DAVID KIRKPATRICK, THE FACEBOOK EFFECT: THE INSIDE STORY OF THE COMPANY

THAT IS CONNECTING THE WORLD 254 (2010) (quoting Mark Zuckerberg stating, “In a lot of

ways Facebook is more like a government than a traditional company. We have this large

community of people, and more than other technology companies we’re really setting

policies.”).

109. See, e.g., Press Release, Dave Yost, Ohio Att’y Gen., AG Yost Files Landmark

Lawsuit to Declare Google a Public Utility (June 8, 2021),

https://www.ohioattorneygeneral.gov/Media/News-Releases/June-2021/AG-Yost-Files-

Landmark-Lawsuit-to-Declare-Google-a [https://perma.cc/NSW6-2TTZ].

110. See Brief of Internet Law Scholars as Amici Curiae Supporting Respondents,

Gonzalez v. Google LLC, 598 U.S. 617 (2023) (No. 21-1333) (signed by 19 Internet law

scholars).

111. See Mitnick, supra note 92, at 826; see also Huang, supra note 92, at 694–95.

112. Huang, supra note 92, at 694–95. Elizabeth Anderson also writes about a study of a

Swiss town that was offered compensation to serve as a potential site for a nuclear waste

facility. She suggests that when the government offered compensation, it regarded the

residents as property owners rather than citizens, with the consequence that residents were less

likely to accept the facility. Anderson, supra note 88, at 197.

113. See Lawrence Lessig, The Regulation of Social Meaning, 62 U.

CHI. L. REV. 943,

1030 (1995).

114. See D

AHRENDORF, supra note 26, at 42–43; Rehbinder, supra note 94, at 953; Dan-

Cohen, supra note 77, at 1233; Anderson, supra note 88, at 193.

115. See Frank Fagan, Systemic Social Media Regulation, 16 D

UKE L. & TECH. REV. 393,

394 (2018).

392239-ILJ 99-4_Text.indd 98392239-ILJ 99-4_Text.indd 98 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1149

no doubt resist being designated a common carrier (indeed, it has),

116

and it is

unlikely to comply voluntarily with its associated behavioral norms (e.g.,

nondiscrimination) because those norms would impair Google’s profits. That is,

Google refuses to internalize its role as a common carrier. Legal sanctions become

crucial—they make the role mandatory by imposing social norms on Google with

the threat of profit-reducing monetary penalties for non-compliance.

117

Legal definitions of social roles serve an additional function for law: they define

the scope of possible lawmaking and legal claims on a particular subject.

118

How law

regulates the sale of sex, for instance, would differ dramatically depending on

whether it regards people who sell sex as “workers” or “prostitutes.” If law adopts

the perspective that those who sell sex are workers, labor rights—like fair wages,

safe working conditions, and, potentially, protected opportunities to organize—are

likely to follow.

119

But if law perceives those who sell sex as prostitutes, it would

condemn the very same acts, prohibit individuals from selling sex, and impose

penalties for violating the prohibition.

120

Associated Press v. United States provides

another helpful example.

121

In that case, the Supreme Court carefully reasoned that

when Associated Press (AP) members entered an agreement not to share their stories

with nonmembers, they acted as businesses engaged in an unreasonable restraint of

trade, not as the press disseminating news for the benefit of democratic society.

122

Legal sanctions against AP members’ agreement to withhold news stories from non-

members flowed from the Court’s perception of AP members’ role as “businesses.”

Law and social roles are mutually constitutive. That is to say, legal definitions of

social roles require public buy-in to be salient and to drive future reform. A group of

political scientists based in the University of Zurich who studied Twitter discourse

around Trump’s ban write:

To rise to the political agenda, a given issue must first be construed as

politically salient and specific arguments put forward as to how and why

it might warrant policy intervention. Therefore, how political actors

frame content moderation may impact the kinds of solutions proposed.

For example, if content moderation is primarily framed as a violation of

free speech policy-makers might be more hesitant to implement strict

116. See State v. Google LLC, No. 21-CV-H-06-0274, 2022 WL 1818648, at *8–9 (Ohio

C.P. May 24, 2022).

117. See D

AHRENDORF, supra note 26, at 42–43.

118. See J.M. Balkin, Understanding Legal Understanding: The Legal Subject and the

Problem of Legal Coherence, 103 Y

ALE L.J. 105, 121–23 (1993).

119. See LaLa B Holston-Zannell, Sex Work is Real Work, and It’s Time to Treat It That

Way, ACLU (June 10, 2020), https://www.aclu.org/news/lgbtq-rights/sex-work-is-real-work-

and-its-time-to-treat-it-that-way/ [https://perma.cc/JL8T-MEDT]; Kate Andrias, The New

Labor Law, 126 Y

ALE L.J. 2 (2016).

120. See, e.g., N.Y.

PENAL LAW § 230.00 (2014),

https://www.nysenate.gov/legislation/laws/PEN/230.00 [https://perma.cc/S6BE-LF5X]. But

see N

EV. REV. STAT. 201.354 (2022), https://www.leg.state.nv.us/nrs/nrs-201.html

[https://perma.cc/LW73-SUA4] (“It is unlawful for a customer to engage in prostitution or

solicitation therefor, except in a licensed house of prostitution.”).

121. 326 U.S. 1 (1945).

122. Id. at 19–20.

392239-ILJ 99-4_Text.indd 99392239-ILJ 99-4_Text.indd 99 5/29/24 10:49 AM5/29/24 10:49 AM

1150 INDIANA LAW JOURNAL [Vol. 99:1131

regulation on platforms’ rules around hate speech, misinformation and

sensitive content.

123

Aileen Nielsen explains that it’s important to understand whether platform users

perceive platforms as governing their speech, because their perception “can be

particularly influential as to what is prioritized in policymaking.”

124

In other words,

if the public sees speech platforms as speech governors and users as speakers and

listeners, they might resist laws that reflect a divergent perception of the speech

platform-user relationship, like common carriage regulation. As Citron and Richards

put it, “Legal rules and policies affecting free expression must take into account the

structures upon which they operate. Legal rules do not operate in a vacuum, and

different rules will operate differently in different structures.”

125

Content moderation

regulation’s legal role-scripting is poised to influence societal expectations about

speech platforms’ appropriate behaviors. If the public accepts, rejects, or alters legal

role-scripts, it will generate headwinds or tailwinds for possible future reform.

Some may feel uneasy at the prospect of law constructing social roles because

social roles typically stem from interpersonal interactions. Douek asks, “Should rules

be expressive and aspirational, especially as rules themselves can set and shape social

norms, or should they reflect the bounds of what is possible in a particular moment .

. . [?]”

126

Top-down prescription of speech platforms’ social role might seem

paternalistic, but at this moment it is unavoidable. Regardless of whether law should

be expressive and aspirational, law in practice shapes social roles whenever it

categorizes those it regulates—whether it redefines existing social roles or creates

them altogether. As Part IV examines in greater depth, current efforts to regulate

content moderation operate on this level. In the face of deep disagreement as to

whether speech platforms are speech governors, common carriers, or publishers,

legal decisions that assert speech platforms should be one of those roles will carry

special force.

III.

CDA SECTION 230 AS A LEGAL ROLE-SCRIPTING FAILURE

Public discourse on content moderation and scholarly discourse on its regulation

reflect that the role of “speech platform” is unsettled and deeply contested. The

previous Part asserted societal uncertainty as to speech platforms’ role positions

current content moderation regulatory reforms to mediate between multiple,

competing accounts of what it means to be a “speech platform” and to behave well

in that capacity.

123. Meysam Alizadeh, Fabrizio Gilardi, Emma Hoes, K. Jonathan Klüser, Mael Kubli &

Nahema Marchal, Content Moderation as a Political Issue: The Twitter Discourse Around

Trump’s Ban 6 (Dep’t Pol. Sci., Univ. Zurich, Working Paper, 2021),

https://fabriziogilardi.org/resources/papers/content-moderation-twitter.pdf

[https://perma.cc/VTF4-9ECM].

124. Aileen Nielsen, The Rights and Wrongs of Folk Beliefs About Speech: Implications

for Content Moderation, 27 UCLA

J.L. & TECH. 118, 130 (2022).

125. Citron et al., supra note 36, at 1381.

126. Douek, Content Moderation as Systems Thinking, supra note 19, at 551.

392239-ILJ 99-4_Text.indd 100392239-ILJ 99-4_Text.indd 100 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1151

This Part examines, given the role theory architecture Part II built, how law

contributed to current discord on speech platforms’ social role. In short, Section 230

of the Communications Decency Act was a legal role-scripting failure. Lawmakers

were not attuned to how the CDA constructed social roles and the statute, as written,

equivocates on its characterization of whom it regulates and whom it serves. As a

result, courts have been left to apply the statute’s liability shields without a firm

understanding of what content moderation norms ought to merit immunity, as

scholars puzzle over what exactly speech platforms are for the purpose of content

moderation regulation.

A. Section 230 as Written and Intended

There are many excellent, extensive accounts of Section 230 and its history and

so this Article presents only a truncated version.

127

Professor Adam Candeub writes

that Congress passed Section 230 in an effort to control pornography and other non-

family-friendly content online.

128

Section 230 aimed to empower parents to control

their children’s access to online content, in part by overruling the New York case,

Stratton Oakmont Inc. v. Prodigy Services Co..

129

In that case, the court held Prodigy

(which managed online bulletin boards) was liable as a “publisher” of statements

made on those bulletin boards.

130

The case “created a Hobson’s choice for . . . content

moderation: either moderate content and face liability for all posts . . . , or don’t

moderate and have posts filled with obscenity or naked images.”

131

To encourage

providers to clean up the internet, Congress passed two liability shields:

Protection for “Good Samaritan” blocking and screening of offensive

material

(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).

132

127. See, e.g., Candeub, supra note 20, at 142–60; Sylvain, supra note 24, at 231–58.

128. See Candeub, supra note 20, at 142.

129. 1995 WL 323710 (N.Y. Sup. Ct. May 24, 1995); see Candeub, supra note 20, at 142.

130. See 1995 WL 323710, at *1 (N.Y. Sup. Ct. May 24, 1995).

131. Candeub, supra note 20, at 142.

132. 47 U.S.C. § 230(c) (1996).

392239-ILJ 99-4_Text.indd 101392239-ILJ 99-4_Text.indd 101 5/29/24 10:49 AM5/29/24 10:49 AM

1152 INDIANA LAW JOURNAL [Vol. 99:1131

Professor Olivier Sylvain, examining the legislative history, explains the bill’s

drafters “proposed the bill as an alternative to direct government restrictions on

speech, believing that ‘parents and families are better suited to guard the portals of

cyberspace and protect our children than our Government bureaucrats.’”

133

The

congressional report accompanying the bill underscores this motivation “to overrule

the Prodigy opinion . . . to further the ‘important federal policy of empowering

parents to determine the content of communications their children receive.’”

134

The core of Section 230(c)(1) is to protect providers from being held liable as

“publishers,” as the New York court previously ruled.

135

Sylvain explains that

Congress’s reference to “publishers” invokes defamation law, in which a publisher

is as liable for publishing defamatory material as its author.

136

Defamation law treats

“distributors” a shade differently. Newsstands, libraries, bookstores, and the like may

only face liability if they actually know the original statement they disseminate is

defamatory.

137

“The [distributor’s] affirmative decision to disseminate the

defamatory statement . . . is a direct cause of . . . harm to the injured party’s

reputation.”

138

In Zeran v. AOL, Inc., the first appellate decision interpreting this

liability shield, the Fourth Circuit reasoned distributor liability effectively treats

distributors as publishers and, by extension, Section 230 must immunize providers

from distributor liability as well.

139

B. Whom Does Section 230 Regulate and Whom Does It Serve?

Though Section 230(c)(1) aimed to counteract publisher liability, it is not clear

from the face of the liability shield whether the law conceived of interactive service

providers as publishers—in other words, whether it viewed providers as playing the

role of “publisher.” The statute’s role choices are far less clear when considering its

prefatory language as well. Before the liability shields, the statute provides:

(a) Findings

The Congress finds the following:

(1) The rapidly developing array of Internet and other interactive

computer services available to individual Americans represent an

extraordinary advance in the availability of educational and

informational resources to our citizens.

(2) These services offer users a great degree of control over the

information that they receive, as well as the potential for even greater

control in the future as technology develops.

133. Sylvain, supra note 24, at 238 (quoting 141 CONG. REC. H8460 (1995) (statement of

Rep. Wyden)).

134. Id. at 239.

135. But see id. (“[T]he history only suggests that Congress did not want the Prodigy

court’s unforgiving interpretation of the republication rule in the online setting to stand.”).

136. Id. at 232–33.

137. Armijo, supra note 45, at 1216.

138. Id. at 1217.

139. 129 F.3d 327, 332 (4th Cir. 1997); see Candeub, supra note 20, at 148–49.

392239-ILJ 99-4_Text.indd 102392239-ILJ 99-4_Text.indd 102 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1153

(3) The Internet and other interactive computer services offer a forum for

a true diversity of political discourse, unique opportunities for cultural

development, and myriad avenues for intellectual activity.

(4) The Internet and other interactive computer services have flourished,

to the benefit of all Americans, with a minimum of government

regulation.

(5) Increasingly Americans are relying on interactive media for a variety

of political, educational, cultural, and entertainment services.

(b) Policy

It is the policy of the United States—

(1) to promote the continued development of the Internet and other

interactive computer services and other interactive media;

(2) to preserve the vibrant and competitive free market that presently

exists for the Internet and other interactive computer services, unfettered

by Federal or State regulation;

(3) to encourage the development of technologies which maximize user

control over what information is received by individuals, families, and

schools who use the Internet and other interactive computer services;

(4) to remove disincentives for the development and utilization of

blocking and filtering technologies that empower parents to restrict their

children’s access to objectionable or inappropriate online material; and

(5) to ensure vigorous enforcement of Federal criminal laws to deter and

punish trafficking in obscenity, stalking, and harassment by means of

computer.

140

With the exception of Section 230(a)(5), it is difficult to see how any of Congress’s

findings or policy articulators relate to providers as “publishers.” Rather, these

provisions point to a number of different social roles the law might support or resist,

without committing to any in particular.

For the most part, the findings section seems to conceive of service providers as

potential collateral government censors and Internet users as citizens. Its emphasis

on providers offering a “forum for a true diversity of political discourse”

141

that

flourishes “to the benefit of all Americans, with a minimum of government

regulation,”

142

suggests Congress took note that Internet users engage online as

citizens in public discourse. It seems Congress was keen to prevent service providers

with the technical ability to control that discourse from becoming censors, to serve

users’ interests as citizens. So, the CDA likely sought to resist a collateral censor-

citizen relationship for service providers and users.

The policy section, on the other hand, seems much more focused on service

providers as businesses and Internet users as consumers. With goals like promoting

the Internet’s development, preserving a free market, encouraging new technologies,

and maximizing user controls,

143

Congress portrays service providers as businesses

and Internet users as consumers in a free market burgeoning with technological

development. Internet users as consumers are primarily interested in choice: “control

140. 47 U.S.C. § 230(a)–(b).

141. Id. § 230(a)(3).

142. Id. § 230(a)(4).

143. Id. § 230(b)(1)–(3).

392239-ILJ 99-4_Text.indd 103392239-ILJ 99-4_Text.indd 103 5/29/24 10:49 AM5/29/24 10:49 AM

1154 INDIANA LAW JOURNAL [Vol. 99:1131

over [what] information . . . they receive.”

144

Service providers, as businesses, pursue

profit motives, but they are hampered by the specter of liability—the liability shields

that follow thus remove financial disincentives to develop blocking and filtering

technologies. Perhaps the CDA enshrines a business-consumer relationship.

The CDA’s since-stricken prohibition on transmitting offensive material to a

minor

145

adds another potential set of roles into the mix: service providers as

broadcasters and Internet users as democratic listeners. That provision criminalized

the “knowing” transmission of “obscene or indecent” messages to any recipient

under 18 years of age.

146

Communications law has a long history of regulating

broadcasters based on an understanding of their relationship to listeners as citizens

of a democracy.

147

Though broadcasters are often private businesses, they are duty

bound to serve listeners’ interests, which include ensuring their programs comport

with important social mores, such as limitations on minors’ access to obscene

content.

148

The CDA as drafted would have transposed this role-relationship on

service providers and Internet users. But, in Reno v. American Civil Liberties Union,

the Supreme Court not only held the provision was unconstitutional, but also

reasoned the Internet is characteristically unlike broadcasting,

149

implying a

broadcaster-listener relationship might be categorically inapplicable to content

moderation.

The liability shields that follow are disjointed from these prefatory sections.

Professor Olivier Sylvain notes the liability shields are auspiciously labeled as

144. Id. § 230(b)(3).

145. 47 U.S.C. § 223(a)(1)(B)(ii) (“§ 223. Obscene or harassing telephone calls in the

District of Columbia or in interstate or foreign communications. (a) . . . Whoever—(1) in

interstate or foreign communications— . . . (B) by means of a telecommunications device

knowingly— . . . (ii) initiates the transmission of, any comment, request, suggestion, proposal,

image, or other communication which is obscene or indecent, knowing that the recipient of

the communication is under 18 years of age, regardless of whether the maker of such

communication placed the call or initiated the communication; . . . shall be fined under title

18 or imprisoned not more than two years, or both.”). The Supreme Court held 47 U.S.C. §

223(a)(1)(B)(ii) unconstitutional in Reno v. ACLU because it violated the First Amendment.

521 U.S. 844, 849 (1997). The Court found the law’s limits on “indecent” and “patently

offensive” were overly vague. Id. at 871–74.

146. 47 U.S.C. § 223(a)(1)(B)(ii).

147. See Syracuse Peace Council v. FCC, 867 F.2d 654, 658 (D.C. Cir. 1989) (explaining

that both the fairness doctrine and its retraction encompass the First Amendment goals of

“stimulating fair, balanced, and diverse treatment of controversial issues (either by single

stations or by the media in the aggregate); of minimizing any chilling effect that may flow

from governmental requirements that one sort of presentation be matched by another; and of

minimizing the risks of abuse and other adverse effects that may flow from having

governmental officials sit in judgment on editorial decisions”); see generally Red Lion Broad.

Co. v. FCC, 395 U.S. 367 (1969) (describing how the fairness doctrine serves listeners’

interests in hearing a broad range of competing viewpoints); Turner Broad. Sys., Inc. v. FCC,

512 U.S. 622 (1994) (reasoning that FCC “must-carry” rules serve Congress’s objective of

providing free television programming to the 40 percent of Americans who lack cable

television).

148. 18 U.S.C. § 1464; 47 U.S.C. § 303.

149. 521 U.S. 844, 868–70 (1997).

392239-ILJ 99-4_Text.indd 104392239-ILJ 99-4_Text.indd 104 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1155

protections for “Good Samaritans.” The label calls up the biblical role of the Good

Samaritan, as well as state statutes that have historically protected individuals from

liability when they attempt to render others aid.

150

Yet, neither shield seems to have

this role in mind.

The first liability shield, Section 230(c)(1), prohibits providers from being treated

as publishers of third-party content, based on an understanding of publishers’

liability for defamation.

151

Defamation (as well as a number of other dignitary torts)

makes certain digressions from societal standards of respectful behavior legally

actionable.

152

It compels publishers to treat people with respect by not disseminating

false and damaging statements about them.

153

The statute’s bar on providers “be[ing]

treated” as publishers thus suggest providers might act like publishers, but it would

be unreasonable to hold them to the behavioral expectations society normally applies

to publishers.

154

The second liability shield makes no mention of publishers. It insulates providers

from liability for their good faith actions to restrict access to material the provider

deems obscene or objectionable.

155

This seems to imagine service providers as

administrative speech governors, like the Federal Communications Commission

(FCC). The FCC is charged by Congress to regulate the wires and waves in the public

interest, and it is accountable to citizens through administrative procedures.

156

It aims

to serve citizens’ interests as democratic listeners, which have historically included

limits on access to obscenity.

157

As for the first shield, the statute seems to invite

service providers to act in this capacity, but it resists societal expectations of public

accountability.

C. Role Equivocation, Role Confusion, and Content Moderation as “Business”

Practice

Citron and Richards assert Section 230 was meant to encourage speech platforms

to monitor and filter the content they host.

158

Sylvain contends Congress sought to

encourage speech platforms “to be passive conduits that facilitate end users’

communications.”

159

The Fifth Circuit, in NetChoice, L.L.C. v. Paxton, found

150. See Sylvain, supra note 24, at 213.

151. 47 U.S.C. § 230(c)(1); see also supra notes 135–136 and accompanying text.

152. Robert Post, The Social Foundations of Defamation Law: Reputation and the

Constitution, 74 C

ALIF. L. REV. 691, 707–19 (1986) (explaining that logically and normatively

defamation causes of action penalize behaviors that disrespect certain social norms).

153. Id.

154. See 47 U.S.C. § 230(c)(1).

155. 47 U.S.C. § 230(c)(2).

156. 47 U.S.C. §§ 307(a), 309(a), (h).

157. See FCC v. Pacifica Found., 438 U.S. 726, 727 (1978) (explaining Congress intended

to give the FCC the power to regulate obscene, indecent, or profane language in exercising its

licensing authority in the public interest).

158. See Citron et al., supra note 36, at 1380.

159. Sylvain, supra note 24, at 219; see also Candeub, supra note 20, at 146 (“[Section

230(c)(1)] treats internet platforms as conduits, such as the telephone or telegraph

companies.”).

392239-ILJ 99-4_Text.indd 105392239-ILJ 99-4_Text.indd 105 5/29/24 10:49 AM5/29/24 10:49 AM

1156 INDIANA LAW JOURNAL [Vol. 99:1131

Section 230 instructs courts not to treat platforms as publishers of user content based

on “Congress’s factual judgment about the role of online platforms”; that is, that

“[they] are not acting as speakers or publishers when they host user-submitted

content.”

160

Who is correct?

The previous section examined in-depth Congress’s handling—perhaps, more

aptly, mishandling—of Section 230’s role constructions. Congress passed Section

230 mainly as a vehicle for more parental control over the material children could

access online, yet it didn’t have a clear image of whom it was regulating (then called

“interactive computer services,” now speech platforms) or the public it served. The

result is a statute that seems to recognize a range of possible roles platforms and their

users might play, without binding platforms to any role-based duties or endowing

users with any role-based rights.

Different perspectives on appropriate content moderation practices talk past each

other because they place speech platforms in different social roles that are often at

cross-purposes. Policymakers that support a common carriage approach, and others

that would impose liability on platforms for failing to take down certain kinds of user

content,

161

cannot see the merits of one another’s position. Each presupposes a

different role-relationship—in this example, common carrier-subscriber and speech

governor-listener.

Section 230’s role equivocation put courts in the unenviable position of

interpreting when to apply Section 230’s liability shields without a firm basis of the

behavioral norms immunity intends to encourage. In the years since the statute’s

passage, courts have immunized an ever-widening scope of speech platforms’

conduct from suit.

162

As Professor Citron and Benjamin Wittes write: “Platforms

have been protected from liability even though they republished content knowing it

might violate the law, encouraged users to post illegal content, changed their design

and policies for the purpose of enabling illegal activity, or sold dangerous

products.”

163

Sylvain adds:

[U]nder current law, a social media company cannot be held responsible

for allowing a user to post compromising private photographs of his ex-

girlfriend publicly. A search engine cannot be called to task under law

for displaying the advertisements of third parties that sell copyrighted

ringtones. An online advertising service is under no legal obligation to

remove posts that encourage the sex trafficking of minors.

164

Most recently, Section 230 came for New York’s privacy statute. New York law

prohibits the use of a person’s name or likeness for advertising or trade purposes

160. 49 F.4th 439, 468 (5th Cir. 2022).

161. Compare infra Part IV.A, with Part IV.B (revealing how different underlying views

about the character of content moderation support dramatically different regulations of content

moderation).

162. See infra notes 163–164 and accompanying text.

163. Danielle Keats Citron & Benjamin Wittes, The Internet Will Not Break: Denying Bad

Samaritans § 230 Immunity, 86 F

ORDHAM L. REV. 401, 408 (2017) (footnote call numbers

omitted).

164. Sylvain, supra note 24, at 217.

392239-ILJ 99-4_Text.indd 106392239-ILJ 99-4_Text.indd 106 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1157

without their consent.

165

Patty Ratermann, a model, sued Amazon, Walmart, and

other online retailers for publishing her photo to advertise a third party’s skincare

product without first obtaining her consent. The Southern District of New York, in

Ratermann v. Pierre Fabre USA, Inc., held Section 230 immunizes the retailers’

publication of her photo because it was the skincare company, a third-party seller on

those sites, who produced the offending content.

166

Though it is troubling for Section

230 to chip away at New York’s only privacy protection, this development is

unsurprising given Section 230(c)(1)’s dissonant role construction. As Part III.B

explains, Section 230(c)(1) suggests speech platforms might act like publishers, but

they shouldn’t be held to publishers’ behavioral norms, which include privacy

standards.

167

The notion that plaintiffs might attempt to hold speech platforms accountable as

“publishers” has led courts and litigants to invest tremendous effort in ascertaining

what it means to be a publisher. The Fourth Circuit in Zeran initiated this focus by

holding “lawsuits seeking to hold a service provider liable for its exercise of a

publisher’s traditional editorial functions—such as deciding whether to publish,

withdraw, postpone or alter content—are barred.”

168

Since then, courts deciding

whether to immunize speech platforms’ content moderation practices from suit

typically frame the issue as whether the content moderation practice amounts to a

publisher’s “traditional editorial functions.” For instance, in Levitt v. Yelp!, Inc., the

Northern District of California held Yelp’s manipulation of review pages, in

violation of its representations to users, was immunized from state unfair or

fraudulent business practices law because the act was “editorial,” even if it was

malicious.

169

As Part III.D examines in greater depth, the Supreme Court in Gonzalez

passed on the chance to decide whether platforms’ algorithmic content

recommendation systems are part of a publisher’s traditional editorial functions or

are the platforms’ own speech.

170

If the Court had held algorithmic content

recommendation constitutes a publisher’s editorial function, it would have

perpetuated Section 230’s dissonant treatment of platforms as publishers (i.e., that

they may act like publishers, but the public shouldn’t expect them to comply with

publishers’ behavioral norms). Alternatively, holding algorithmic content

recommendation falls outside a publisher’s editorial function would have disclaimed

speech platforms’ likeness to publishers, but otherwise left their role uncertain.

Scholars and lawmakers attempt to resolve Section 230’s role indeterminacy by

debating who has the correct perspective of what speech platforms are as a matter of

empirical reality.

171

But, in the background, speech platforms put into practice their

self-understanding that they moderate content as “businesses.” Professor Kate

165. N.Y. CIV. RIGHTS LAW §§ 50–51 (2023),

https://www.nysenate.gov/legislation/laws/CVR/A5 [https://perma.cc/LW3G-6J9W].

166. 651 F. Supp. 3d 657, 665-68 (S.D.N.Y. Jan. 17, 2023).

167. See supra notes 151-154 and accompanying text.

168. Zeran v. AOL, Inc., 129 F.3d 327, 330 (4th Cir. 1997).

169. Nos. C–10–1321, C–10–2351, 2011 WL 5079526, at *6–7 (N.D. Cal. Oct. 26, 2011),

aff’d, 765 F.3d 1123 (9th Cir. 2014).

170. Gonzalez v. Google LLC, 598 U.S. 617 (2023); see infra notes 173-185 and

accompanying text.

171. See supra notes 47-73 and accompanying text.

392239-ILJ 99-4_Text.indd 107392239-ILJ 99-4_Text.indd 107 5/29/24 10:49 AM5/29/24 10:49 AM

1158 INDIANA LAW JOURNAL [Vol. 99:1131

Klonick writes, “[T]he primary reason companies take down obscene and violent

material is the threat that allowing such material poses to potential profits based in

advertising revenue. Platforms’ ‘sense of the bottom-line benefits of addressing hate

speech can be shaped by consumers’ . . . expectations.’”

172

That is to say, speech

platforms typically engage in content moderation to the extent it coincides with their

profit motives and, simultaneously, they regard their users as consumers who reveal

their preferences in the amount of time they engage with the platform.

173

If speech platforms are just “businesses” that moderate content if and when it

serves their profit interests, the walls close in on the possibility of any content

moderation regulation. That is because the commercial speech doctrine, as it has

developed since the 1970s, affords increasing First Amendment protection to

business-to-consumer speech.

174

If framed as business-to-consumer speech, the

practice of content moderation would be insulated from governmental regulation.

175

In Virginia State Board of Pharmacy v. Virginia Citizens Consumer Council, the

Court held commercial speech warrants some protection from governmental

regulation, on the grounds “[i]t is a matter of public interest that [purchasing]

decisions, in the aggregate, be intelligent and well informed. To this end, the free

flow of commercial information is indispensable.”

176

“[I]f it is indispensable to the

proper allocation of resources in a free enterprise system, it is also indispensable to

the formation of intelligent opinions as to how that system ought to be regulated or

altered.”

177

The Court applied this constitutionalized business-consumer relationship to

digital information transactions in Sorrell v. IMS Health Inc.

178

In that case, the Court

held a Vermont law that forbade the sale or use of records containing doctors’

prescribing practices for marketing purposes violated the First Amendment.

179

reasoned, “The State may not burden the speech of others in order to tilt public debate

in a preferred direction. ‘The commercial marketplace, like other spheres of our

social and cultural life, provides a forum where ideas and information flourish.’”

180

Whereas earlier the Court understood commercial speech to serve consumers’

interest in the free flow of information, the Court now seems to conceive that

businesses have a constitutionally protected interest in speaking—and, by extension,

in engaging in commercial transactions that support their ability to speak.

181

172. Kate Klonick, The New Governors: The People, Rules, and Processes Governing

Online Speech, 131 H

ARV. L. REV. 1598, 1627 (2018) (quoting Danielle Keats Citron & Helen

Norton, Intermediaries and Hate Speech: Fostering Digital Citizenship for Our Information

Age, 91 B.U.

L. REV. 1435, 1454 n. 113 (2011)).

173. See id.

174. See infra notes 176–181 and accompanying text.

175. See Armijo, supra note 45, at 1228.

176. Va. State Bd. of Pharmacy v. Va. Citizens Consumer Council, 425 U.S. 748, 765

(1976).

177. Id.

178. 564 U.S. 552 (2011).

179. Id. at 557.

180. Id. at 578–79 (quoting Edenfield v. Fane, 507 U.S. 761, 767 (1993)).

181. See Amy Kapczynski, The Lochnerized First Amendment and the FDA: Toward a

More Democratic Political Economy, 118 C

OLUM. L. REV. ONLINE 179, 200–01 (2018);

392239-ILJ 99-4_Text.indd 108392239-ILJ 99-4_Text.indd 108 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1159

Some scholars and courts assert “content moderation policies are protected

speech”

182

based on an implied business-consumer framing, but that is not a foregone

conclusion. Though this Part described the CDA as a legal role-scripting failure,

there is some upside to the role indeterminacy it engendered. Whether speech

platforms ought to be regarded as “businesses” is not settled as a matter of law or

societal expectation. The range of roles scholars and courts invoke when they

evaluate platforms’ liability for content moderation reveals as much. The First

Amendment is an obstacle to content moderation regulation, but that obstacle is more

like a hurdle than a brick wall. The next Part of this Article examines two current

content moderation regulatory reforms—law’s second take at role-scripting—to

direct scholars’ and lawmakers’ attention to important questions about whether they

establish desirable social roles for speech platforms and the public.

D. Gonzalez v. Google’s Residual Uncertainty on Speech Platforms as Publishers

In Gonzalez v. Google, the Supreme Court had the opportunity to consider

whether speech platforms act as publishers when they recommend (algorithmically

or otherwise) one user’s content to another.

183

The family of Nohemi Gonzalez, an

American killed in an ISIS terrorist attack in Paris, sued Google, Twitter, and

Facebook for violating the Anti-Terrorism Act (ATA).

184

They alleged the speech

platforms aided and abetted terrorism by recommending terrorism recruitment

content to their users.

185

The Ninth Circuit held the conduct was immune from suit under Section 230

because “the duty that the plaintiff[s] allege[] the defendant violated derives from

the defendant’s status or conduct as a ‘publisher or speaker’” of another user’s

Jedediah Purdy, Neoliberal Constitutionalism: Lochnerism for a New Economy, 77 L. &

CONTEMP. PROBS. 195, 200–01 (2014).

182. Armijo, supra note 45, at 1228 (emphasis omitted). See also Eric Goldman & Jess

Miers, Online Account Terminations/Content Removals and the Benefits of Internet Services

Enforcing their House Rules, 1 J.

OF FREE SPEECH L. 191, 192 (2021) (explaining that court

decisions reflect that platforms have the right to terminate user accounts); Johnson v. Twitter,

No. 18cecg00078 (Cal. Sup. Ct. June 6, 2018) (dismissing plaintiff’s claims Twitter’s

suspension of his account due to his political viewpoint violated the California Constitution,

the Unruh Act, and the UCL, based on application of Section 230 immunity and the First

Amendment); Taylor v. Twitter, A154973 (Cal. Apt. Ct. Aug. 17, 2018) (same); Federal

Agency of News LLC v. Facebook, Inc., 395 F. Supp. 3d 1295, 1309 (N.D. Cal. 2019)

(“Courts have rejected the notion that private corporations providing services via the internet

are public fora for purposes of the First Amendment.”); Tulsi Now, Inc. v. Google, LLC, No.

2:19-cv-06444, 2020 WL 4353686 (C.D. Cal. Mar. 3, 2020) (rejecting plaintiff’s claim Google

violated her First Amendment rights partly on the ground that its company self-regulation is

no way equivalent to government regulation); Wilson v. Twitter, No. 3:20-cv-00054, 2020

WL 3410349 (S.D.W.V. May 1, 2020) (dismissing plaintiff’s claim Twitter violated his First

Amendment rights based partly on the premise “Twitter is a publicly traded, multi-national

corporation and not an arm of federal or state government”).

183. Petition for Writ of Certiorari, at *i, Gonzalez v. Google LLC, No. 21-1333, 2022 WL

1050223 (Apr. 4, 2022).

184. See Gonzalez v. Google LLC, 2 F.4th 871, 880 (9th Cir. 2021); 18 U.S.C. § 2333.

185. Gonzalez, 2 F.4th at 881.

392239-ILJ 99-4_Text.indd 109392239-ILJ 99-4_Text.indd 109 5/29/24 10:49 AM5/29/24 10:49 AM

1160 INDIANA LAW JOURNAL [Vol. 99:1131

content.

186

Two of the three judges on the panel (one concurring and the other

partially concurring and partially dissenting) asserted that if not bound by precedent,

they would “hold that the term ‘publisher’ under section 230 reaches only traditional

activities of publication and distribution—such as deciding whether to publish,

withdraw, or alter content—and does not include activities that promote or

recommend content or connect content users to each other.”

187

The effect of the

dissent’s preferred course would be to deny speech platforms Section 230 immunity

for their content recommendation.

188

The Supreme Court granted Gonzalez’s petition for certiorari but ultimately

declined to opine on Section 230’s application to the platforms’ content

recommendation.

189

In its three-page per curiam opinion, it instead vacated the Ninth

Circuit’s judgment and remanded to the lower court because Petitioners likely failed

to state a claim under the ATA.

190

The merits of Gonzalez hinged on an understanding of the societal expectations

placed on publishers—their “traditional editorial functions”—and whether a speech

platform’s content recommendations fall within them. Petitioners asserted “it strains

the English language to say that in . . . recommending . . . writings to users . . . [one]

is acting as the publisher of . . . information provided by another information content

provider.”

191

Recommending books is not inherently a publisher’s function because

anyone can recommend a book. Respondents, on the other hand, equated user content

recommendation to a decision to “display[] . . . information” traditionally associated

with publishers.

192

Amici added a range of interpretations of what publishers do and

whether those expectations include content recommendation.

193

The social role of

“publisher” was at the very center of the Gonzalez case.

186. Id. at 891 (quoting Barnes v. Yahoo!, Inc., 570 F.3d 1096, 1102 (9th Cir. 2009)).

187. Id. at 913 (Berzon, J., concurring); see also id. at 921–21 (Gould, J., concurring in

part and dissenting in part) (“In short, I do not believe that Section 230 wholly immunizes a

social media company’s role as a channel of communication for terrorists in their recruiting

campaigns and as an intensifier of the violent and hatred-filled messages they convey.”).

188. Id. at 925 (Gould, J., concurring in part and dissenting in part) (“I would hold that

Section 230 does not immunize Google from liability for its content-generating algorithms

insofar as they develop a message to ISIS-interested users. The same reason supports lack of

immunity for the other Defendant social media companies’ use of their own algorithms,

procedures, users, friends, or other means to deliver similar content from ISIS to the users of

the social media.”).

189. Gonzalez v. Google LLC, 598 U.S. 617, 622 (2023) (per curiam). The Court reached

this holding in reliance on Twitter v. Taamneh, 598 U.S. 471 (2023), which it decided on the

same day. Id. at 478.

190. Id.

191. Petition for Writ of Certiorari, at *27, Gonzalez v. Google LLC, No. 21-1333, 2022

WL 1050223 (Apr. 4, 2022) (quoting Force v. Facebook, 934 F.3d 53, 76–77 (2d Cir. 2019)

(Katzmann, J., dissenting)) (internal quotations omitted) (alteration in original) (emphasis in

original).

192. See Brief in Opposition at *20, Gonzalez v. Google LLC, No. 21-1333, 2022 WL

2533118 (Jul. 5, 2022) (alteration omitted) (quoting id. at *31a).

193. Brief of Amici Curiae Common Sense Media and Frances Haugen in Support of

Petitioners, at 3, Gonzalez v. Google LLC, No. 21-1333 (arguing Google isn’t a publisher

because it “also monitors and tracks users’ online activities across websites and computer

392239-ILJ 99-4_Text.indd 110392239-ILJ 99-4_Text.indd 110 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1161

There are two ways the Court could have decided, each scripting speech

platforms’ social role differently. First, the Court could have decided content

recommendation isn’t a behavioral norm attributed to “publishers” and so speech

platforms could face liability for the recommendations they make. As Judge Berzon

put it in his concurrence, “The actions of the social network algorithms—assessing

a user’s prior posts, friends, or viewing habits to recommend new content and

connections—are more analogous to the actions of a direct marketer, matchmaker,

or recruiter than to those of a publisher.”

194

Second, the Court could have decided

content recommendation falls within a “publisher’s” behavioral norms and, on that

basis, immunized speech platforms from most claims their recommendations violate

the law. This would accord with the Chamber of Commerce’s view that publishers

inherently recommend whatever content they choose to publish.

195

Either way, the Court would not have rectified Section 230’s role-scripting

failure. The first possible decision would have disclaimed any affinity between

speech platforms and publishers (much in the way the Court in Reno v. ACLU

reasoned interactive computer services are characteristically unlike broadcasters).

196

But that decision would otherwise leave speech platforms’ role indeterminate. Other

laws, perhaps the PACT Act or state-level common carriage laws, might then have

supplied role-based behavioral expectations (as well as sanctions for deviations) to

fill the gap.

197

The second possible decision—finding speech platforms act as

publishers when they recommend content—would have entrenched the role

dissonance Section 230(c)(1) first generated. Part III.B explains that the CDA’s first

liability shield suggests speech platforms might act like publishers, but it would be

unreasonable for individuals to hold them to the behavioral norms that typically

apply to publishers. There isn’t a clear path forward for law that follows this role

construction, whether it’s the application of existing law or newly promulgated laws

that don’t amend or repeal Section 230(c)(1).

devices to make predictions about the content users want to view.”); Brief of Fairplay as

Amicus Curiae in Support of Petitioner Reynaldo Gonzalez et al., at 26–27, Gonzalez v.

Google LLC, No. 21-1333 (explaining courts misconceive the term “publisher” when they

“immuniz[e] companies for their own conduct in designing social media algorithms, products,

and environments that affirmatively harm children.”) (emphasis omitted); Amicus Brief of

Senator Josh Hawley in Support of Petitioners, at 4–8, Gonzalez v. Google LLC, No. 21-1333

(explaining distributor liability is distinct from publisher liability and it is not subject to

Section 230 immunity); Brief for the States of Tennessee, et al. as Amici Curiae in Support of

Petitioners, at 10–15, Gonzalez v. Google LLC, No. 21-1333 (arguing in favor of a narrow

view of “publisher” immunity); Brief of Amicus Curiae of Professor Eric Goldman in Support

of Respondent, passim, Gonzalez v. Google LLC, No. 21-1333 (construing speech platforms

as “online publishers”); Brief of Washington Legal Foundation as Amicus Curiae Supporting

Respondent, at 15–16, Gonzalez v. Google LLC, No. 21-1333 (contending Congress intended

“publisher” to have its broad, everyday meaning); Brief of TechFreedom as Amicus Curiae in

Support of Respondent, at 10, Gonzalez v. Google LLC, No. 21-1333 (“A platform, as it

disseminates third-party content, is a publisher.”).

194. Gonzalez, 2 F.4th at 915 (Berzon, J., concurring).

195. Brief of Amicus Curiae Chamber of Commerce of the United States of America in

Support of Respondent, at 14, Gonzalez v. Google, 598 U.S. 617 (2023).

196. 521 U.S. 844, 868–70 (1997).

197. See supra Parts IV.A and IV.B.

392239-ILJ 99-4_Text.indd 111392239-ILJ 99-4_Text.indd 111 5/29/24 10:49 AM5/29/24 10:49 AM

1162 INDIANA LAW JOURNAL [Vol. 99:1131

In the wake of the Supreme Court’s (in)decision in Gonzalez v. Google, federal

and state lawmakers are left to articulate their own visions of speech platforms’ social

roles and regulate them accordingly.

198

The Part that follows evaluates two current

content moderation regulatory reforms in terms of the social roles they construct for

speech platforms and the public.

IV.

CONTENT MODERATION REGULATIONS’ ROLE SCRIPTS

How law talks about speech platforms matters, now more than ever. Public and

scholarly discourse on content moderation and its regulation reflect that there is no

settled understanding of what the role of “speech platform” entails, what constitutes

“good” or “normal” content moderation, and what interests individuals might

reasonably have in a content moderation relationship. Deep divisions and heated

arguments over who has the correct perspective owe in part to Section 230’s failure

to articulate a cohesive vision of whom it regulates and the public it serves. Section

230 fostered uncertainty as to whether speech platforms should act as publishers,

speech governors, or common carriers and serve public interests associated with each

of these roles. Section 230 was a role-scripting failure, but lawmakers—especially

the Supreme Court—now have another opportunity to script speech platforms’ social

role as they moderate users’ content. The roles law chooses today may help the public

and legal authorities understand whether Elon Musk’s content moderation decisions

are despotic acts of censorship, discrimination on the basis of users’ viewpoints, or

editorial judgments.

This Part examines two current content moderation regulatory reforms in terms

of the social roles they would construct for speech platforms and individuals: the

PACT Act and Protect Speech Act in Congress

199

and Texas and Florida common

carriage laws,

200

as well as the state laws’ judicial review. It surfaces their role

constructions by interrogating the messages they express about whom they regulate

and whom they protect. These include statements about the governed relationship’s

legal attributes (e.g., control, dependency) and parties’ relevant attributes and

interests (e.g., participation, respect), as well as role characteristics implied from the

rights, duties, behavioral constraints, and entitlements that attach to those who meet

expressed attributes. It then examines how each reform’s social roles would guide

behavior beyond explicit legal requirements and direct future legal reform.

This Part does not aim to settle debate over whether speech platforms, as they

exist today, are more “like” speech governors, common carriers, or cable operators,

etc. Rather, it starts from the position that the decisions lawmakers make today about

how to regulate content moderation articulate a vision of what speech platforms

should be. Content moderation regulations are expressive and norm-setting. This Part

instead raises questions lawmakers, scholars, and the public ought to answer

198. See supra note 189.

199. Protect Speech Act, H.R. 3827, 117th Cong. (2021); PACT Act, S. 4066, 116th Cong.

(2020).

200. H.B. 20, 87th Leg., 2d Called Sess. (Tex. 2021); S.B. 7072, 123d Leg., Reg. Sess.

(Fla. 2021).

392239-ILJ 99-4_Text.indd 112392239-ILJ 99-4_Text.indd 112 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1163

regarding whether these reforms support desirable role-based behavioral norms

within the content moderation relationship.

201

A. The PACT Act & The Protect Speech Act: Speech Platforms as Speech

Governors

The speech governance narrative that features prominently in scholarly discourse

and public accounts of Musk’s Twitter takeover is perhaps best reflected in two bills

before Congress: the PACT Act and the Protect Speech Act. These bills cast speech

platforms as administrative speech governors and individuals as speakers and

listeners in a democratic society.

202

The Protect Speech Act, for one, would require speech platforms that wish to use

Section 230 as a defense to state publicly, in detail, their content moderation rules,

comply with those rules, and provide a rationale and an opportunity for the person to

respond when they restrict access to a person’s content.

203

The PACT Act requires

the same and also requires speech platforms to create a complaint filing system that

allows individuals to file and track complaints and appeals to platforms’ content

moderation decisions.

204

These proposals would afford what Citron calls

“technological due process,” which provides individuals with processes analogous

to those required of a government whenever it deprives a person of life, liberty, or

property.

205

Those processes include providing individuals with notice, the

opportunity to be heard, and an impartial tribunal.

206

201. A word of caution is due. It is not possible for law to predict with great clarity the

direction social meaning will take as law seeks to influence norms, nor is it possible for any

human to do the same. This Part evaluates the societal expectations each reform’s choice of

social roles support if society comes to accept and apply these roles to the content moderation

relationship. This hypothetical posture is necessary because society may reject law’s choice

of social roles or redefine them substantially in unpredictable ways. Still, it is helpful to

consider the societal expectations each role-relationship supports because of law’s greater

likelihood to influence social roles for relationships in which they are not yet settled.

202. These bills seem to draw from scholars’ work construing content moderation as a

form of speech governance. See Klonick, supra note 172, at 1603 (arguing content moderation

governs users’ speech); Jack M. Balkin, The First Amendment in the Second Gilded Age, 66

UFFALO L. REV. 979, 997 (2018) (agreeing with Kate Klonick); Jack M. Balkin, Free Speech

Is a Triangle, 118 C

OLUM. L. REV. 2011, 2021–22 (2019) (explaining online platforms govern

their spaces in multiple ways, including content moderation) [hereinafter Balkin, Triangle];

Alina Tugend, Barred from Facebook, and Wondering Why, N.Y.

TIMES, Sept. 20, 2014, at

B5 (describing the effect Facebook account termination has on people’s lives); T

ARLETON

GILLESPIE, CUSTODIANS OF THE INTERNET (2018) (describing online intermediaries as

exercising a governance function on the internet); Rebecca Tushnet, Content Moderation in

an Age of Extremes, 10 C

ASE W. RES. J.L. TECH. & INTERNET 1, 14 (2019) (acknowledging

speech platforms’ likeness to governors).

203. See Protect Speech Act, H.R. 3827, 117th Cong. (2021).

204. See PACT Act, S. 4066, 116th Cong. (2020).

205. See Danielle Keats Citron, Technological Due Process, 85 W

ASH. U. L. REV. 1249,

1305–13 (2008).

206. See id. at 1278–88 (citing Mullane v. Cent. Hanover Bank, 339 U.S. 306, 313 (1950)).

392239-ILJ 99-4_Text.indd 113392239-ILJ 99-4_Text.indd 113 5/29/24 10:49 AM5/29/24 10:49 AM

1164 INDIANA LAW JOURNAL [Vol. 99:1131

Not all speech governors are treated—or are expected to act—like governments.

Churches, private universities, and the military govern speech in different ways and

to different ends. Expectations about how each should be responsive (or non-

responsive) to those they govern depend on the source of their legitimacy.

207

Speech

is governed by churches to serve a higher power; by private universities in pursuit of

knowledge creation and, more recently, a welcoming environment for individuals of

diverse backgrounds; and by the military to maintain good order and discipline in

service of national security.

208

Law does not require due process of these

organizations, nor could it reasonably because they regulate speech within

relationships that do not require democratic legitimacy.

209

The PACT Act and the Protect Speech Act thus seem to construe speech platforms

and their users to be in a quasi-governmental relationship—a relationship compatible

with democratic processes. That is because they react to speech platforms’

governance of speech in the digital public sphere.

210

Consider, for example,

Facebook’s update to its content moderation policy to ban misinformation about

COVID-19 and the consequent removal of twenty million pieces of rule-violating

content, or its long-standing, highly specific restrictions of certain forms of nudity

and sexual content.

211

When Facebook promulgates and enforces these rules, it

defines how individuals may engage in public discourse on political matters and in

207. See generally ROBERT C. POST, DEMOCRACY, EXPERTISE, AND ACADEMIC FREEDOM

(2012) (explaining that different institutions, such as religious organizations, universities, and

the military, regulate speech differently because of their different underlying purposes and

objectives).

208. See generally id.

209. See generally id.

210. The kinds of harm these bills address, such as a lack of accountability and procedure,

reflect their roots in academic discourse on the digital public sphere. See Balkin, Triangle,

supra note 202, at 2012, 2041 (“Social media companies and search engines have social and

moral obligations to the public because they perform three connected public services. First,

they facilitate public participation in art, politics, and culture. Second, they organize public

conversation so that people can easily find and communicate with each other. Third,

they curate public opinion by providing individualized feeds and search results, and by

enforcing civility norms through their terms-of-service obligations and community

guidelines.”); Jack M. Balkin, Free Speech in the Algorithmic Society: Big Data, Private

Governance, and New School Speech Regulation, 51 U.C.

DAVIS L. REV. 1149, 1197 (2018)

(“The more that end-users view businesses as governors, or as special-purpose sovereigns, the

more end-users will expect—and demand—that these companies should conform to the basic

obligations of governors towards those they govern. These obligations include procedural

fairness in handling complaints and applying sanctions, notice, transparency, reasoned

explanations, consistency, and conformity to rule of law values—the “law” in this case being

the publicly stated norms and policies of the company.”).

211. COVID-19 and Vaccine Policy Updates & Protections, F

ACEBOOK HELP CTR.,

https://www.facebook.com/help/230764881494641 [https://perma.cc/2NBH-

D7MQ?type=image]; Monika Bickert, How We’re Taking Action Against Vaccine

Misinformation Superspreaders, M

ETA (Aug. 18, 2021),

https://about.fb.com/news/2021/08/taking-action-against-vaccine-misinformation-

superspreaders/ [perma.cc/B6MM-TFXR]; Adult Nudity and Sexual Activity, Objectionable

Content, Community Standards, M

ETA, https://transparency.fb.com/policies/community-

standards/adult-nudity-sexual-activity/ [https://perma.cc/BDJ2-W9JU].

392239-ILJ 99-4_Text.indd 114392239-ILJ 99-4_Text.indd 114 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1165

cultural development more broadly. Its content moderation policies and practices

shape the digital public sphere much like public law would.

The proposals thus cast platforms as administrative speech governors, which are

expected to govern public speech democratically in the public interest. Accordingly,

they should make decisions about permissible and impermissible speech, but those

decisions must meet citizens’ needs as speakers and listeners in a democratic society.

Speakers’ and listeners’ interests differ. Speakers are expected to be interested

primarily in individual and collective autonomy, i.e., the ability to express their own

views in an overall effort to shape law that is responsive to the collective public

will.

212

Listeners, on the other hand, desire access to speech on a wide array of

diverse viewpoints, but also a body of common facts to ground the public debate,

and, historically, limits on public speech that defies public morals, like obscenity.

213

The role-relationship between speech governors and democratic

speakers/listeners derives from the historic part the FCC played regulating speech in

the public sphere. Congress mandated the FCC to promulgate broadcasting rules and

regulations “specifically directed to consider the demands of the public.”

214

Historically, the FCC took up this mandate by promulgating the “fairness doctrine,”

which required broadcasters to cover controversial public issues and afford speakers

the opportunity to present contrasting viewpoints,

215

afford a certain amount of

airtime to educational programming,

216

and limit when indecent and profane content

could air.

217

Moreover, the FCC, like other administrative agencies, must follow

procedural requirements, including public notice and comment on proposed rules.

218

These obligations ensure the FCC, as a speech governor, fulfills the requirements of

democratic accountability.

Placing speech platforms and individuals in a speech governance relationship

helps explain a number of contemporary problems, like platforms’ circulation of

212. See generally DANIELLE KEATS CITRON, HATE CRIMES IN CYBERSPACE (2014)

(discussing speakers’ autonomy interests in the context of how law can regulate online hate

crimes); R

OBERT C. POST, CONSTITUTIONAL DOMAINS: DEMOCRACY, COMMUNITY,

MANAGEMENT (1995) (explaining the normative foundations of First Amendment protections

for speakers); P

OST, supra note 207 (describing the importance of individual and collective

autonomy for democratic functioning).

213. See Jane Bambauer, Is Data Speech?, 66 S

TAN. L. REV. 57, 59 (2014) (describing

listeners’ interests in receiving “raw facts”); C.

EDWIN BAKER, MEDIA, MARKETS, AND

DEMOCRACY (2001) (examining liberal and republican democratic theories on the public’s

need to access viewpoints that accord with their own and information that establishes common

ground among diverse individuals); P

OST, supra note 207 (explaining the importance of

expertise in providing the public with facts that allow them to pursue a common good); FCC

v. Pacifica Found., 438 U.S. 726, 727 (1978) (situating the FCC’s obscenity restrictions in its

authority to regulate the airways in the public interest).

214. See Red Lion Broad. Co. v. FCC, 395 U.S. 367, 380 (1969) (citing 47 U.S.C. §§

307(a), 309(a)).

215. See Report Concerning General Fairness Doctrine Obligations of Broadcast

Licensees, 102 F.C.C. 2d 143, 146 (1985); Syracuse Peace Council v. FCC, 867 F.2d 654, 655

(D.C. Cir. 1989).

216. See 61 Fed. Reg. 43981 (Aug. 27, 1996) (codified at 47 CFR §§ 73.671, 73.673).

217. 47 CFR § 73.3999.

218. See Administrative Procedure Act, 5 U.S.C. §§ 551–559.

392239-ILJ 99-4_Text.indd 115392239-ILJ 99-4_Text.indd 115 5/29/24 10:49 AM5/29/24 10:49 AM

1166 INDIANA LAW JOURNAL [Vol. 99:1131

misinformation, promotion of politically divisive content, and filter bubbles. Each is

inappropriate because it undermines listeners’ interests in accessing diverse

viewpoints and common facts. Platforms as speech governors also act

inappropriately when they make “black box” content moderation decisions, either by

relying on algorithmic decision-making that obscures the basis for a content

moderation decision or by following “secret” internal rules that diverge from those

made public.

219

Overall, the speech governance relationship directs speech platforms

to moderate content in the public interest (as opposed to platforms’ profit interest) in

a manner that comports with public morals and is responsive to the will of those

governed. It also guides individuals to treat speech platforms as rightful governors—

to accept the legitimacy of their content moderation authority and voice their dissent

to platforms’ policies and decisions.

The specific platform behavioral expectations this relationship could support may

fluctuate dramatically as society evolves and politics unfold. First, speakers’ and

listeners’ interests point in different directions, and they can easily come into

conflict. One can imagine a Facebook user who wants to post her belief that 2020

presidential election workers stuffed fraudulent ballots into suitcases and another

user who wants that post to be removed as misinformation. The first user acts as a

speaker, interested in expressing her political views. The second acts as a listener,

concerned with their access to shared facts. Platforms as speech governors are

expected to act in the public interest. That means they will have to make difficult

decisions about what the public interest requires that some will undoubtedly disagree

with. Second, behavioral expectations down the line might follow a number of tracks.

Perhaps individuals will expect platforms to follow an updated “fairness rule,” such

that they must include differing viewpoints on controversial matters. In a sense,

common carriage proposals already point in that direction.

220

Or they might expect

platforms to be interoperable with smaller rivals, to support public access to diverse

moderation regimes. Though substantive expectations might vary widely, they share

a baseline understanding that platforms have the authority to regulate speech in

general.

Reforms that place platforms and individuals into a speech-governance

relationship may start out fairly limited in scope, along the lines of the Protect Speech

Act and the PACT Act. These laws would only tweak content moderation at the

margins. They largely morph what prominent speech platforms already do into a set

of legal requirements.

221

However, law that adopts this relationship could justify both

broader and more extensive regulation down the line. As Professor Rebecca Tushnet

puts it, “if what we are talking about are truly governance structures, why should

citizens of a democracy accept anything other than actual democracy, either

representative or otherwise, in the regulation of these spaces?”

222

A speech

219. Klonick, supra note 172, at 1631–34.

220. See infra Part IV.B.

221. Compare Klonick, supra note 172, at 1630–49 (demonstrating that large speech

platforms, including but not limited to Facebook, already engage in numerous, bureaucratic

internal and external governance mechanisms when moderating content, likely largely

compliant with the Acts’ requirements), with supra Part I.A (describing the Acts’

requirements).

222. Tushnet, supra note 202, at 16.

392239-ILJ 99-4_Text.indd 116392239-ILJ 99-4_Text.indd 116 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1167

governance relationship could justify an explicit legal obligation on platforms to

moderate content in the public interest. It could also support reforms that would

impose a wider range of democratic governance features—such as an Administrative

Procedure Act with which platforms must comply if they seek Section 230’s

coverage

223

—or even more dramatic forms of public control, like binding user-

initiated ballots on content moderation policies.

224

The speech governance relationship departs significantly from the business-

consumer relationship that frames platforms’ approach to content moderation. It

engenders the expectation that platforms regulate speech as a public service and,

accordingly, they should make content moderation decisions in the public interest,

rather than in their profit interest, and be subject to some form of public control. It

would be unsurprising for public-interest content moderation decisions to require

platforms to sacrifice some profit; decisions about politically divisive issues will

alienate some users and advertisers. But the cost to platforms (even if ultimately

passed on to consumers) is not a primary concern of regulation that responds to a

speech governance relationship. Instead, it aims to secure a public sphere that can

sustain democratic government.

On the surface, additional due process might sound inherently good. It seems to

empower individuals, in a general sense, by affording them rights against an

otherwise more powerful relationship counterpart. But due process rights are not a

given for all power-asymmetrical relationships: they must be justified based on the

roles participants play within that relationship. The PACT Act and the Protect Speech

Act thus depend on casting speech platforms as speech governors and individuals as

democratic speakers and listeners. They send a message that speech platforms ought

to be speech governors and individuals should regard them as such. Whether these

social roles are good is far from inherent. Rather, lawmakers and the public should

consider whether it’s desirable for platforms to engage in governance and whether

they should be charged with difficult decisions about how to moderate content in the

public interest. Likewise, they should scrutinize whether it’s desirable for individuals

to respect platforms’ sovereignty over public discourse or to participate in content

moderation rulemaking.

B. Texas & Florida Platform Common Carriage Laws: Speech Platforms as

Common Carriers (Maybe?)

Concern over speech platforms’ ability to limit individuals’ ability to speak on

their platforms has animated conservative lawmakers, legal scholars, and even

Justice Clarence Thomas to suggest law should regulate speech platforms as

common carriers.

225

Recently promulgated laws in Texas and Florida attempt to do

223. See Van Loo, supra note 49, at 875–86 (proposing in detail a set of “Platform Federal

Rules” akin to the Administrative Procedure Act).

224. See Bloch-Wehba, supra note 19, at 75–76.

225. See Matthew Feeney, Opinion, Are Social Media Companies Common Carriers?,

CATO

INST. (May 24, 2021), https://www.cato.org/blog/are-social-media-companies-

common-carriers [https://perma.cc/G5RY-HGAS] (“There is a widespread mistrust of

household‐name social media companies among many American conservatives, who allege

that Big Tech is institutionally biased against conservatives and seeking to stifle their

392239-ILJ 99-4_Text.indd 117392239-ILJ 99-4_Text.indd 117 5/29/24 10:49 AM5/29/24 10:49 AM

1168 INDIANA LAW JOURNAL [Vol. 99:1131

as much. Texas’s H.B. 20 prohibits social media companies from banning users

based on their “viewpoint,” whereas Florida’s S.B. 7072 penalizes social media

platforms for removing or obstructing access to content posted by or about political

candidates.

226

But, as written, the statutes are murky on whether they cast speech

platforms as discriminatory common carriers or censorial speech governors.

The two statutes differ on finer points, but H.B. 20 demonstrates the role

confusion that plagues both. The Texas legislature writes, in its findings, “social

media platforms function as common carriers, are affected with a public interest, are

central public forums for public debate, and have enjoyed governmental support in

the United States.”

227

These findings generally accord with the criteria Yoo identifies

as historically relevant in designating an entity as a common carrier.

228

But the

statute’s core conduct proscription bears little connection to common carriage. It

provides: “A social media platform may not censor a user . . . based on: (1) the

viewpoint of the user or another person; (2) the viewpoint represented in the user’s

expression or another person’s expression; or (3) a user’s geographic location in this

state . . ..”

229

The statute also requires platforms to maintain complaint-and-appeal

systems for aggrieved users.

230

These behavioral prohibitions and mandates speak

far more to a relationship of governance demanding democratic legitimacy.

Censorship is generally problematic when done by a state

231

and a complaint-and-

appeal system mirrors the due process rights the PACT Act provides.

A common carriage relationship, by contrast, is concerned chiefly with

nondiscriminatory service on fair terms. Common carriage duties trace to the

emergence of railroads as a dominant mode of transportation of both people and

goods.

232

At the common law, railroads were bound by a “duty to serve at just and

content.”); David Yost, Let’s Make Google a Public Good, N.Y. TIMES (July 7, 2021),

https://www.nytimes.com/2021/07/07/opinion/google-utility-antitrust-technology.html

[https://perma.cc/B5J9-FDF8] (describing Attorney General Yost’s reasoning for declaring

Google a public utility); Biden v. Knight First Am. Inst. at Columbia Univ., 141 S. Ct. 1220,

1222–23 (2021) (Thomas, J., concurring) (describing how law may set limits on social media

companies’ right to exclude users by implementing common carriage regulation).

226. H.B. 20, 87th Leg., 2d Called Sess. (Tex. 2021); S.B. 7072, 2021 Leg., Reg. Sess.

(Fla. 2021). Other efforts to regulate speech platforms as common carriers include Ohio

Attorney General Dave Yost’s declaration, via lawsuit, that Google is a common carrier, see

Press Release, Dave Yost, Ohio Att’y Gen., AG Yost Files Landmark Lawsuit to Declare

Google a Public Utility (June 8, 2021), https://www.ohioattorneygeneral.gov/Media/News-

Releases/June-2021/AG-Yost-Files-Landmark-Lawsuit-to-Declare-Google-a

[https://perma.cc/5TPR-VR8H], and the 21st Century Free Speech Act in Congress, see 21st

Century FREE Speech Act, S. 1384, 117th Cong. (2021). The bill would require platforms, as

common carriers, to provide their service to anyone without discriminating against individual

users or classes of users, or on the basis of political or religious affiliation or region. Id.

227. Tex. H.B. 20 § 1(3).

228. See Yoo, supra note 20 and accompanying text.

229. Tex. H.B. 20 § 143A.002.

230. See id. §§ 120.101–04.

231. See generally C

ENSORSHIP AND SILENCING: PRACTICES OF CULTURAL REGULATION

(Robert C. Post ed., 1998) (describing various cultural practices of expressive control).

232. See James B. Speta, A Common Carrier Approach to Internet Interconnection, 54

ED. COMM. L.J. 225, 253–58 (2002).

392239-ILJ 99-4_Text.indd 118392239-ILJ 99-4_Text.indd 118 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1169

reasonable prices,” not just in the case of a monopoly, but because railroads “were

affected by [the] ‘public interest’ in that all members of the public needed to rely on

their availability.”

233

Public interest concerns also animated the Interstate Commerce

Act of 1887 (ICA), the first federal legislation imposing common carrier duties,

which imposed requirements of just and reasonable rates and conditions of service,

nondiscrimination, and tariff filing, and a mild prohibition on price discrimination.

234

Later amendments to the ICA required railroads to interconnect their lines as well.

235

The Communications Act of 1934 extended common carriage duties to companies

engaged in interstate communication by wire, such as telephone and telegraph

companies.

236

If these companies “hold[] [themselves] out indiscriminately to a class

of persons for service,” they have a duty to serve the public.

237

Much like the

railroads before them, telephone companies’ duty to serve first required just and

reasonable rates and nondiscrimination and later expanded to interconnection with

other carriers.

238

The common carriage components of the Texas and Florida statutes attempt to

translate the historic relationship between telecommunications carriers and

subscribers to the platform content moderation context. It casts speech platforms as

common carriers that principally serve to transport information between individuals,

much like telephones and, by some accounts, the internet’s backbone. Consider, for

instance, Facebook’s display of a user’s Friends’ posts in the user’s Newsfeed, or

Google’s transmission of search results that respond to a user query. A common

carriage relationship construes a platform’s information transport services as

essential because the platform is either the sole provider in its industry or one of

few.

239

Though Facebook’s and Google’s respective market shares are currently in

dispute,

240

it would be uncontroversial to say there are relatively few alternatives to

each.

241

The logic of common carriage regulation suggests that because platforms provide

essential information transport services, they are bound by a duty to serve.

242

That

is, they are expected to serve all who seek their service, on just and reasonable terms,

233. Id. at 256–57.

234. Id. at 258–59.

235. Id. at 260.

236. The ICA brought telecommunications into its purview prior to the Communications

Act’s promulgation. Id. at 261–63.

237. Id. at 264.

238. Id. at 263–64.

239. See Nikolas Guggenberger, Essential Platforms, 24 S

TAN. TECH. L. REV. 237, 252–

87 (2021).

240. See Order Granting Motion to Dismiss, FTC v. Facebook, Inc., 560 F. Supp. 3d 1

(D.D.C. 2021).

241. Popular estimates suggest Facebook, Twitter, Pinterest, and YouTube account for

roughly 98% of worldwide social media use and Google, Bing, Baidu, and Yahoo! comprise

roughly 98% of worldwide search engine use. See Social Media Stats Worldwide,

STATCOUNTER (Nov. 2021), https://gs.statcounter.com/social-media-stats

[https://perma.cc/CM5U-ABVV]; Search Engine Market Share Worldwide,

STATCOUNTER

(Nov. 2021), https://gs.statcounter.com/search-engine-market-share [https://perma.cc/WZV8-

3CEH].

242. Feeney, supra note 225.

392239-ILJ 99-4_Text.indd 119392239-ILJ 99-4_Text.indd 119 5/29/24 10:49 AM5/29/24 10:49 AM

1170 INDIANA LAW JOURNAL [Vol. 99:1131

without discrimination. Moreover, they are expected to not regulate the morals or

conduct of those they serve: “Public utilities and common carriers [such as telephone

companies] are not the censors of public or private morals, nor are they authorized

or required to . . . regulate the public or private conduct of those who seek service at

their hands.”

243

Some may push back on the notion that “common carrier” is a social role at all

and not just a legal category. But, as Volokh points out, email systems are not treated

as common carriers as a matter of law, and they are technically able to screen

messages based on the viewpoints they contain.

244

Even so, it would surprise people

if their email provider screened emails in this way, because “an e-mail system’s ‘role

in transmitting e-mail is akin to that of a telephone company, which one neither wants

nor expects to superintend the content of its subscribers’ conversations.’”

245

Individuals commonly regard email providers as playing the role of “common

carrier”—that is to say, they expect providers to fulfill a common carrier’s behavior

norms—though law does not require it.

A common carriage relationship casts individuals as subscribers interested in

receiving platforms’ information transport services (i.e., having their messages and

desired content faithfully transmitted) for fair prices and other terms of service.

Subscribers’ interest is subtly different than a typical consumer’s interest in price

and quality: the former is concerned with fairness. Fair prices and terms of service

do not hinge on market prices and quality levels, but rather the notion that essential

transportation providers must serve the public interest. Fair prices and terms are thus

tied to the particular public interests at stake given the provider’s service offering.

Whereas it might be unfair for a railroad to provide preferential transport to corn

over wheat, in the case of speech platforms, it might be unfair to promote user content

that supports military withdrawal from the Middle East over user content that

supports military expansion.

The common carriage relationship also depicts individuals as dependent on

platforms because they provide information transport services that are necessary for

daily life.

246

One can imagine how difficult it would be to connect with others online

without social media platforms like Facebook or find information across webpages

without a search engine like Google. The law’s expectation that individuals are

dependent on platforms helps normalize this dependency.

A common carrier-subscriber relationship helps explain the harm individuals

perceive when they are kicked off a speech platform or their post is removed because

of its content. That is because platforms act inappropriately if they refuse to serve a

particular person or group for reasons that are unfair or discriminatory. What counts

as unfair or discriminatory is open to interpretation, contest, and further elaboration

by society. Evidently, conservatives believe it is unfair for platforms to treat content

differently on the basis of the political or religious viewpoint it expresses.

247

In their

view, notions of nondiscrimination derived from First Amendment doctrine that

243. Volokh, supra note 20, at 386 (quoting Pa. Pubs. Inc. v. Pa. Pub. Util. Comm’n, 36

A.2d 777, 781 (Pa. 1944)).

244. Id. at 386.

245. Id. (quoting Lunney v. Prodigy Servs. Co., 723 N.E.2d 539, 542 (N.Y. 1999)).

246. Guggenberger, supra note 239, at 252–76.

247. See 21st Century FREE Speech Act, S. 1384, 117th Cong. (2021).

392239-ILJ 99-4_Text.indd 120392239-ILJ 99-4_Text.indd 120 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1171

condemns governmental restrictions of public discourse

248

should apply to speech

platforms’ transmission of user content.

249

On the other hand, LGBT activists assert

it is unfair for platforms to take down content that expresses LGBT viewpoints on

the basis it violates rules that restrict sexual material.

250

Their conception of

nondiscrimination derives from platforms’ intermediation of cultural exchange and

keys off civil rights laws that protect progressive norms from orthodox repression.

251

Whether either of these forms of content moderation is unfair or discriminatory will

depend on the rough consensus that emerges from political debate.

Whatever the result, grounding the content moderation relationship in

expectations of fairness, nondiscrimination, and a duty to serve allows concrete

platform behavioral expectations to fluctuate along with society—and expand or

contract accordingly. It also limits platforms’ discretion to determine what qualifies

as fair or discriminatory; platforms are bound by public political will.

If law responds to a common-carriage relationship, a range of reforms may

proceed that support subscribers’ interest in fairness or derive from platforms’ duty

to serve the general public. Law may scrutinize platforms’ terms of service beyond

their content moderation policies. Perhaps it is unfair to monetize individuals’

personal information in exchange for providing information transport services, to

bind them to mandatory arbitration, or to prohibit them from using “web spiders” to

automatically gather information from the platform.

252

Platforms’ duty to serve could

248. See Robert Post, Meiklejohn’s Mistake: Individual Autonomy and the Reform of

Public Discourse, 64 U

NIV. COLO. L. REV. 1109, 1115–16 (1993) (examining the normative

basis for the First Amendment to prohibit governmental viewpoint-based restrictions on

speech).

249. See Richard Epstein, Should Platforms Be Treated as Common Carriers? It Depends.,

American Enterprise Institute (July 2022), https://platforms.aei.org/should-platforms-be-

treated-as-common-carriers-it-depends/ [https://perma.cc/E33T-MZAU]; Phillip Hamburger,

Setting the Record Straight on Reining in Big Tech, New Civil Liberties Alliance (Aug. 5,

2021), https://nclalegal.org/2021/08/setting-the-record-straight-on-reining-in-big-tech/

[https://perma.cc/YE7J-39QS]. That’s to say, conservatives don’t derive their claim from

existing nondiscrimination law, which (for the most part) doesn’t consider political ideology

a “protected class.” See Craig R. Senn, Ending Political Discrimination in the Workplace

Ending Political Discrimination in the Workplace, 87 M

O. L. REV. 365, 373-405 (2022)

(cataloguing disparities at the federal and state level, and between public and private sector

employees, in terms of whether employment laws protect them from discrimination on the

basis of their political viewpoints).

250. Greg Bensinger & Reed Albergotti, YouTube Discriminates Against LGBT Content

by Unfairly Culling It, Suit Alleges, W

ASH. POST (Aug. 14, 2019, 4:21 PM),

https://www.washingtonpost.com/technology/2019/08/14/youtube-discriminates-against-

lgbt-content-by-unfairly-culling-it-suit-alleges/ [https://perma.cc/GJU2-33XF]; Oliver L.

Haimsen, Daniel Delmonaco, Peipei Nie & Andrea Wegner, Disproportionate Removals and

Differing Content Moderation Experiences for Conservative, Transgender, and Black Social

Media Users: Marginalization and Moderation Gray Areas, at 5–6, Proc. ACM Hum.-

Comput. Interact. 5, CSCW2, Article 466 (Oct. 2021),

https://dl.acm.org/doi/pdf/10.1145/3479610 [https://perma.cc/48RZ-N95P].

251. Fagan, supra note 115, at 394.

252. See, e.g., Thomas E. Kadri, Platforms as Blackacres, 68 UCLA

L. REV. 1184, 1186–

87 & nn. 7, 368, 369 (2022) (explaining how the Computer Fraud and Abuse Act may endow

websites with too much authority to impose civil and criminal liability on entities that use

392239-ILJ 99-4_Text.indd 121392239-ILJ 99-4_Text.indd 121 5/29/24 10:49 AM5/29/24 10:49 AM

1172 INDIANA LAW JOURNAL [Vol. 99:1131

also support mandatory interconnection with other platforms, to the extent

interconnection would afford more of the public access to the platform.

253

The

overriding concerns of a law that responds to a common carriage relationship are

service to all on equal terms coupled with fair prices and conditions of carriage.

The state common carriage laws exhibit confusion about the roles they intend to

apply to speech platforms and individuals. This risks replicating the CDA’s legal

role-scripting failure and perpetuating public confusion about what speech platforms

are and how they should moderate content. Even taken in isolation, the common

carriage aspects of the state laws raise important questions about whether they

support a positive content moderation relationship.

Much like the speech governance relationship the PACT Act would establish, at

the surface common carriage requirements sound appealing. Platforms’ duty to serve

all, on fair terms, and without engaging in discrimination seem to advance the values

of equity and public participation. But the common carriage relationship on which

these duties rely for justification might be less appealing. Should individuals accept

dependence on a few platforms (rather than reject dependence and demand

government action that supports new competition)? What constitutes fair content

moderation terms and who should be empowered to decide? Does a duty to serve

mean platforms must interconnect with one another, even if it comes at a cost to

individuals’ privacy? These are some questions about common carriage’s role-based

behavioral expectations lawmakers and the public ought to consider when deciding

whether to support this type of content moderation reform.

C. The NetChoice Cases: Speech Platforms as Cable Operators

Shortly after the Texas and Florida common carriage laws were promulgated,

NetChoice, an industry group representing social media platforms, brought suit.

254

argued the laws were unconstitutional on the ground they interfere with the

platforms’ exercise of “editorial judgment” through content moderation, which it

contended is speech protected by the First Amendment.

255

The Fifth Circuit upheld

Texas’s law, in part based on the court’s conclusion that platforms’ content

moderation isn’t an exercise of “editorial judgment”

256

protected by the First

automated tools, like web spiders, to gather information from them).

253. See, e.g., Fiona Scott Morton, Opinion, Why ‘Breaking Up’ Big Tech Probably Won’t

Work, W

ASH. POST (July 16, 2019, 2:41 P.M.),

https://www.washingtonpost.com/opinions/2019/07/16/break-up-facebook-there-are-

smarter-ways-rein-big-tech/ [https://perma.cc/XV3Z-3VC7] (arguing law should mandate

platform interconnection rather than break up big technology companies).

254. See NetChoice, LLC v. Moody, 546 F. Supp. 3d 1082, 1084 (N.D. Fla. 2021), aff’d

in part, vacated in part, remanded sub nom. NetChoice, LLC v. Att’y Gen., Fla., 34 F.4th

1196 (11th Cir. 2022); NetChoice, LLC v. Paxton, 573 F. Supp. 3d 1092, 1101 (W.D. Tex.

2021), vacated and remanded sub nom. NetChoice, LLC v. Paxton, 49 F.4th 439 (5th Cir.

2022).

255. NetChoice v. Paxton, 573 F. Supp. 3d at 1103; NetChoice v. Moody, 546 F. Supp. 3d

at 1085.

256. The two opinions use the phrases “editorial judgment” and “editorial discretion”

interchangeably.

392239-ILJ 99-4_Text.indd 122392239-ILJ 99-4_Text.indd 122 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1173

Amendment.

257

By contrast, the Eleventh Circuit rejected most of Florida’s law

based on its conclusion that speech platforms exercise “editorial judgment” much

like cable operators.

258

Both courts evaluated the pivotal question of whether platforms’ content

moderation involves protected “editorial judgment” by evaluating platforms’

similarity to other roles, like newspapers, parades, and cable operators.

259

The mode

of reasoning went: the Supreme Court established that X role(s) engage in “editorial

judgment” when they transmit speech, so if platforms are sufficiently like X role(s),

their content moderation must also be a form of “editorial judgment.”

260

The Fifth

Circuit dispensed with the three roles, finding they share common characteristics

(i.e., “carefully curat[ing]” their “choice of material”; and “select[ing] . . . content

before that content is hosted, published or disseminated”)

261

and common behavioral

expectations (i.e., “accept[ing] reputational and legal responsibility for the content

[they] edit[]”)

262

that platforms don’t fulfill. The Eleventh Circuit also reasoned

whether platforms have a protectable speech interest by evaluating their likeness to

those social roles, but it reached a different conclusion: “[S]ocial-media platforms

should be treated more like cable operators, which retain their First Amendment right

to exercise editorial discretion . . . .”

263

The Eleventh Circuit asserted that the First Amendment protects a “private

entity’s choices about whether, to what extent, and in what manner it will disseminate

speech” because they are “editorial judgments.”

264

The court seems to purport that

the First Amendment protects the dissemination of speech regardless of an entity’s

social role. However, its reasoning conflicts with that assertion—to determine

whether a particular act constitutes protected “editorial judgment,” the court

distinguished between different entities on the basis of their roles. For example, law

schools and shopping malls’ “choices about whether, to what extent, and in what

manner [they] will disseminate speech” do not qualify as “editorial judgments,” but

when those decisions are made by newspapers, parade organizers, or cable operators,

they do.

265

Ultimately, the court regarded social media platforms like cable operators,

257. NetChoice v. Paxton, 49 F.4th at 464–65.

258. NetChoice v. Att’y Gen., Fla., 34 F.4th at 1220.

259. NetChoice v. Paxton, 49 F.4th, at 459–65; NetChoice v. Att’y Gen., Fla., 34 F.4th at

1203–23.

260. See NetChoice v. Paxton, 49 F.4th, at 459–65; NetChoice v. Att’y Gen., Fla., 34 F.4th

at 1203–23.

261. NetChoice v. Paxton, 49 F.4th at 459, 461, 464 (emphasis in original).

262. Id. at 464.

263. NetChoice v. Att’y Gen., Fla., 34 F.4th at 1220.

264. Id. at 1210.

265. Id. at 1210–21.

392239-ILJ 99-4_Text.indd 123392239-ILJ 99-4_Text.indd 123 5/29/24 10:49 AM5/29/24 10:49 AM

1174 INDIANA LAW JOURNAL [Vol. 99:1131

because both are technically unlike broadcast media,

266

and both refuse to carry

certain content “not to propound a particular point of view.”

267

Though the Eleventh Circuit’s conclusion that law should treat platforms like

cable operators saved platforms from much of the Florida law (i.e., the provisions

the court held were “content-based”),

268

its decision to equate platforms to cable

operators is hardly a boon for platforms. The court’s opinion elides the fact that cable

operators are regulated quite heavily on the basis of their social role. In fact, however

paradoxical to the Eleventh Circuit’s decision, the FCC requires cable operators that

allow a “legally qualified candidate for public office . . . to use its facilities” to afford

“equal opportunities to all other candidates for that office to use its facilities,”

without censoring that candidate’s material in any way.

269

The FCC regulates cable operators in myriad ways in line with its view of their

social role; that is, its view that if cable operators hold “undue market power” they

would imperil “the availability of diverse views and information.”

270

On that basis,

the FCC authorizes state governments to engage in subscription fee regulation in

some cases; requires cable operators to carry at least one noncommercial educational

station; requires cable operators to carry all content of the broadcast stations they

transmit (without alteration or deletion); requires cable operators to provide

subscribers certain parental control; specifies subscriber privacy protections; restricts

cable operators’ ownership of programming producers; and caps each cable

operator’s nationwide reach, among other things.

271

The Copyright Act also requires

cable operators to remit compulsory license payments to content providers.

272

In Turner Broadcasting System v. FCC, the Supreme Court upheld FCC “must

carry” rules, which required cable operators to set aside a certain number of channels

for over-the-air broadcasters who elect mandatory carriage.

273

The Court refused to

hold the rules unconstitutional, on the grounds the rules were content neutral and

Congress found that cable operators had “undue market power” that threatened the

viability of free broadcast television.

274

It reached this holding despite its judgment

that “cable operators engage in and transmit speech, and they are entitled to the

protection of . . . the First Amendment.”

275

266. NetChoice v. Att’y Gen., Fla., 34 F.4th at 1220. The Court relies on the Supreme

Court’s decisions in Turner Broad. Sys., Inc. v. FCC, 512 U.S. 622 (1994), and Reno v.

American Civil Liberties Union, 521 U.S. 844 (1997). Turner held cable operators shouldn’t

be subject to less stringent First Amendment scrutiny that applies to broadcasters because

cable lacks the “unique physical limitations” (i.e., the scarcity of frequencies) that

characterizes broadcast television. 512 U.S. at 637–39. The Supreme Court similarly

distinguished the Internet from broadcast media in Reno. 521 U.S. at 844, 868–69.

267. NetChoice v. Att’y Gen., Fla., 34 F.4th at 1213 (quoting Hurley v. Irish-American

Gay, Lesbian & Bisexual Group of Boston, 515 U.S. 557, 575 (1995).

268. See id. at 1226.

269. 47 C.F.R. § 76.205 (2019).

270. Cable Television, FCC (June 15, 2021),

https://www.fcc.gov/media/engineering/cable-television [https://perma.cc/2AGQ-7H8R].

271. Id.

272. 17 U.S.C. § 115 (2018).

273. Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622, 628–32 (1994).

274. See id. at 632–34.

275. Id. at 636–38.

392239-ILJ 99-4_Text.indd 124392239-ILJ 99-4_Text.indd 124 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1175

The Eleventh Circuit’s decision that speech platforms are “like” cable operators

invites policymakers to query whether speech platforms’ “undue market power”

likewise risks the availability of diverse content the public requires. Likeness to cable

operators supports the expectation that speech platforms have a responsibility to

carry diverse content for the listening public’s benefit, and when they accumulate an

excess of market power, law is authorized to the extent it pursues that public interest.

That characterization of speech platforms’ social role enables a range of regulation

affecting content moderation and beyond. It could justify regulation not terribly

different from the Florida “common carriage” law—as for cable operators, a

requirement that if a platform allows a political candidate to transmit over their

services, it must afford other candidates for the same office equal opportunity. It

could support law that restricts content moderation to initial decisions to carry or not

carry a particular user and prohibits later deletion or alteration of that user’s content.

It could also support more stringent privacy protections and potentially compulsory

license payments to the users whose content the platforms profit from.

ONCLUSION

The Article argues the core project of content moderation regulation is legal role-

scripting. Whereas scholarly discourse commonly portrays law as role “taking” in

this space,

276

this Article characterizes law as chiefly role “making.”

277

The decisions

lawmakers make about how to regulate content moderation will likely form the basic

contours of what it means to be a speech platform, how one moderates content

appropriately, and the interests individuals reasonably have when they engage with

speech platforms.

278

The relationship between social roles and law is not a line, with

law merely reflecting existing roles, but a loop. When social roles are unsettled or

uncertain—as in the content moderation relationship—law is especially poised to

influence their preliminary, basic contours.

279

Lawmakers can attempt to script social roles well for the content moderation

relationship by being conscientious about behavioral norms and legal pathways that

flow from their regulatory decisions. The final Part of this Article aimed to assist that

effort by surfacing difficult questions about the desirability of different reforms’ role

scripts and the online speech ecosystem they produce.

280

Examining these contours

of content moderation regulations demands time and attention. This Article asserts it

is time well spent.

It is not a given that current regulatory reforms will rectify the CDA’s failure to

articulate cohesive and consistent roles within the content moderation relationship.

281

Rather, if lawmakers are not conscientious about the social roles content moderation

regulations construct, they might replicate, entrench, or worsen the role uncertainty

the CDA first established.

276. See supra Part I.B.

277. See supra Parts I.C & II.B.

278. See supra Parts I.C & II.B.

279. See supra Part II.B.

280. See supra Part IV.

281. See supra Part III.C.

392239-ILJ 99-4_Text.indd 125392239-ILJ 99-4_Text.indd 125 5/29/24 10:49 AM5/29/24 10:49 AM

1176 INDIANA LAW JOURNAL [Vol. 99:1131

It is equally important to acknowledge that at least five distinct legal authorities

are in the process of articulating competing views of speech platforms’ role.

Congress is considering whether to hold speech platforms accountable as

administrative speech governors;

282

Texas and Florida have attempted to categorize

speech platforms as common carriers;

283

the Eleventh Circuit espoused its view that

speech platforms engage in “editorial judgment” akin to publishers;

284

and the

Supreme Court may wade into the debate on whether speech platforms act as

publishers when it resolves the Fifth and Eleventh Circuits’ split over the

constitutionality of the Texas and Florida laws.

285

There is a strong prospect these legal authorities will simultaneously establish

conflicting roles for speech platforms and the public, and that is a problem. Indeed,

the Fifth and Eleventh Circuits have already clashed on whether speech platforms

should behave like “common carriers,” or like cable operators that exercise “editorial

judgment.”

286

Texas’s common carrier relationship indicates platforms have a duty

to serve all who seek their service, on just and reasonable terms, without

discrimination.

287

The Eleventh Circuit’s publisher relationship authorizes platforms

to make unilateral decisions about how they will moderate content, regardless

whether they serve listeners’ interests.

288

The PACT Act’s speech governance

relationship would exacerbate those competing expectations by proposing platforms

will and should moderate content, but in speakers’ and listeners’ interests and

keeping with due process-like requirements.

289

Sometimes, a common carrier’s reasonable, nondiscriminatory content

moderation might resemble a speech governor’s content moderation in speakers’ and

listeners’ interests and a cable operator’s editorial judgment. For instance, a

moderation rule that allows users to post breastfeeding photos, despite a general

prohibition on nudity, might be reasonable and nondiscriminatory—in that it

comports with prevailing social norms and refrains from prejudicing mothers—and

serve speakers’ interest in expressing their views and listeners’ interest in accessing

a rich and diverse public discourse.

These role-relationships diverge on a deeper level. First, the speech governor

relationship binds platforms to moderate content in the public interest, whereas the

common carriage relationship presumes platforms are bound by a duty to serve all,

and a cable operator’s unchecked editorial judgment shirks any normative binds

whatsoever.

290

Platforms’ content moderation decisions will likely differ on that

basis. The public interest may warrant robust content moderation that doesn’t afford

service to all people or all content; a duty to serve suggests only minimal moderation

will be appropriate; and unchecked editorial judgment could support maximalist or

282. See supra Part IV.A.

283. See supra Part IV.B.

284. See NetChoice v. Att’y Gen., Fla., 34 F.4th at 1220.

285. NetChoice, LLC v. Paxton, No. 22-555 (argued Feb. 26, 2024); Moody v. NetChoice,

LLC, No. 22-277 (argued Feb. 26, 2024).

286. See supra notes 254–63 and accompanying text.

287. See supra Part IV.B.

288. See supra Part IV.C.

289. See supra Part IV.A.

290. See supra Part IV.

392239-ILJ 99-4_Text.indd 126392239-ILJ 99-4_Text.indd 126 5/29/24 10:49 AM5/29/24 10:49 AM

2024] CONTENT MODERATION REGULATION 1177

minimalist moderation or (as Elon Musk has exhibited)

291

unpredictable vacillation

between the two.

Second, the PACT Act affords individuals due process rights when platforms

moderate content because they serve the interests of speakers and listeners in public

discourse.

292

A common carriage regulation wouldn’t make that allowance because

it doesn’t serve subscribers’ interest in receiving information transport services on

fair terms.

293

Individual due process makes even less sense against a cable operator—

if someone disagrees with an editorial decision, they ought to write a letter to the

editor.

294

The role conflict on the horizon presages further uncertainty about speech

platforms’ role within a content moderation relationship. It is difficult to tell how the

mix of federal statute, federal case law, and state statute would direct speech

platforms’ content moderation practices in general, let alone in a way that serves

important social values. Moreover, legally induced role conflict would likely

entrench public disagreement about what speech platforms are and how they should

moderate appropriately. Each side could point to a source of law to bolster the

reasonableness of their position.

It will likely be up to the Supreme Court to resolve this role conflict. It declined

to do so in the Gonzalez case,

295

but the Court still might articulate its perception of

speech platforms’ social role when settles the Fifth Circuit and Eleventh Circuit’s

split on the constitutionality of platform-common carriage laws.

296

Those cases

examined a wide range of roles speech platforms might play—as speech governors,

common carriers, publishers, and others.

297

Whereas in Reno the Court asserted the

Internet is characteristically unlike broadcasting,

298

the Court could now put its

weight behind one cohesive and consistent vision of speech platforms’ social role.

291. See supra notes 1–10 and accompanying text.

292. See supra Part IV.A.

293. See supra Part IV.B.

294. See supra Part IV.C.

295. Gonzalez v. Google LLC, 598 U.S. 617, 622 (2023) (per curiam).

296. See NetChoice, LLC v. Paxton, No. 22-555 (argued Feb. 26, 2024); Moody v.

NetChoice, LLC, No. 22-277 (argued Feb. 26, 2024); NetChoice, LLC v. Paxton, 49 F.4th

439, 445 (5th Cir. 2022); NetChoice, LLC v. Att’y Gen., Fla., 34 F.4th 1196, 1209–10 (11th

Cir. 2022).

297. See NetChoice v. Paxton, 49 F.4th at 455–59; NetChoice v. Att’y Gen., Fla., 34 F.4th

at 1215–18.

298. Reno v. American Civil Liberties Union, 521 U.S. 844, 868–70 (1997).

392239-ILJ 99-4_Text.indd 127392239-ILJ 99-4_Text.indd 127 5/29/24 10:49 AM5/29/24 10:49 AM