Amicus Curiae Brief — Reynaldo Gonzalez, et al., Petitioners v. Google LLC
Supreme Court briefJan 18, 2023
Ask Donna
What actually matters in this document.
Text
No. 21-1333
In the Supreme Court of the United States
REYNALDO GONZALEZ, ET AL.,
PETITIONERS,
V.
GOOGLE LLC,
RESPONDENT.
___________
ON WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
___________
BRIEF OF PUBLIC KNOWLEDGE AS AMICUS CURIAE
IN SUPPORT OF RESPONDENT
___________
JOHN BERGMAYER
ALEX PETROS
HAROLD FELD
Counsel of Record
PUBLIC KNOWLEDGE
1818 N St. NW Suite 410
Washington, DC 20782
(202) 861-0020
hfeld@publicknowledge.org
January 18, 2023
ii
TABLE OF CONTENTS
TABLE OF CONTENTS ....................................................ii
TABLE OF AUTHORITIES ............................................... iv
INTEREST OF AMICUS CURIAE ....................................... 1
SUMMARY OF ARGUMENT .............................................. 2
ARGUMENT ................................................................... 3
I.
YouTube’s Recommendations Are a Form of
Publishing ......................................................... 3
A. Section 230 Precludes Liability Under All
Causes of Action That Seek to Impose Liability
for User Material ................................................ 5
B. Publishing Is Expressive Activity That Often
Conveys a Message of Endorsement ................. 9
C. Congress Enacted Section 230 To Encourage
Providers to Use Their Independent Editorial
Judgment to Moderate Content ....................... 13
D. “Recommendations” Are Not Distinct from
Other Features of Interactive Computer
Services ............................................................. 14
II.
Section 230 Applies to Publishing, Not
Publisher Business Models ............................ 17
A. Section 230 Does Not Prevent Holding Online
Marketplaces Accountable for Selling
Dangerous Products ......................................... 18
B. Section 230 Does Not Prevent State and Local
Business Regulation ......................................... 19
C. Section 230 Does Not Shield the Collection and
Use of Personal Data ........................................ 19
III.
Section 230 Was Intended to Promote
Competition, Free Expression, and User
Control ............................................................ 21
iii
A. Section 230 Protects and Promotes a
Competitive Market ......................................... 21
B. Only Congress Can Update Section 230 to
Better Fulfill Its Purposes in Today’s
Circumstances .................................................. 23
CONCLUSION ............................................................... 26
iv
TABLE OF AUTHORITIES
Cases
Anderson v. TikTok, 2022 U.S. Dist. LEXIS 193841
(E.D. Pa. 2022) ....................................................... 25
Arizona v. Mayorkas, 598 U.S. __ (2022) ................. 24
Banaian v. Bascom, 281 A.3d 975 (2022) .................. 4
Barnes v. Yahoo!, 570 F.3d 1096 (9th Cir. 2009) ... 6, 7
Batzel v. Smith, 333 F.3d 1018 (9th Cir. 2003) ......... 8
Brooke Group v. Brown & Williamson Tobacco, 509
U.S. 209 (1993) ....................................................... 22
Brooks v. Thomson Reuters Corp., No. 21-cv-01418EMC, 2021 U.S. Dist. LEXIS 154093 (N.D. Cal.
Aug. 16, 2021) ........................................................ 20
Doe v. GTE Corp., 347 F.3d 655 (7th Cir. 2003) ........ 6
Dyroff v. Ultimate Software Grp., Inc., 934 F.3d 1093
(9th Cir. 1093) ........................................................ 25
Fair Hous. Council v. Roommates.com, 521 F.3d
1157 (9th Cir. 2008) ........................................... 8, 20
Firth v. State, 98 N.Y.2d 365 (2002)........................... 4
Force v. Facebook, 934 F. 3d 53 (2d Cir. 2019) .... 4, 11
Giuffre v. Dershowitz, 410 F. Supp. 3d 564 (S.D.N.Y.
2019) ......................................................................... 4
Henderson v. Source For Pub. Data, 53 F.4th 110
(4th Cir. 2022). ......................................................... 7
Herrick v. Grindr, 765 F. App’x 586, 590-91 (2d Cir.
2019) ......................................................................... 7
HomeAway.com v. City of Santa Monica, 918 F.3d
676, 684 (9th Cir. 2018) ......................................... 19
In re Phila. Newspapers, 690 F.3d 161 (3d Cir. 2012)
.................................................................................. 6
Maynard v. Snapchat, 357 Ga. App. 496 (Ga. Ct.
App. 2020) .............................................................. 25
v
Miami Herald Pub. Co., Div. of Knight Newspapers.
v. Tornillo, 418 U.S. 241, 258 (1974). ............. 10, 12
Oberdorf v. Amazon.com Inc., 930 F.3d 136 ............ 18
Oja v. United States Army Corps of Engineers, 440
F.3d 1122 (9th Cir. 2006). ....................................... 9
Salyer v. S. Poverty Law Ctr., 701 F. Supp. 2d 912
(W.D. Ky. 2009) .................................................... 4, 5
Verizon Communs. v. Law Offices of Curtis V.
Trinko, 540 U.S. 398 (2004) .................................. 22
Zeran v. America Online, 129 F.3d 327, 330 (4th Cir.
1997) ....................................................................... 12
Statutes
141 Cong Rec H 8460 ................................................ 13
47 U.S.C. § 230 .................................................. passim
PACT Act, S. 797, 117th Cong. (2021) ..................... 24
Platform Integrity Act, H.R. 9695, 118th Cong.
(2022) ...................................................................... 24
S. Rep. No. 104-230 ................................................... 13
Other Authorities
Claire Cain Miller, When Uber and Airbnb Meet the
Real World, NY Times (Oct. 17, 2014),
https://www.nytimes.com/2014/10/19/upshot/whenuber-lyft-and-airbnb-meet-the-real-world.html. .. 18
Congressional Research Service, Section 230: An
Overview (April 7, 2021) ........................................ 15
DuckDuckGo, Measuring the Filter Bubble
(December 4, 2018),
https://spreadprivacy.com/google-filter-bubblestudy ....................................................................... 18
Harold Feld, The Case for the Digital Platform Act:
Market Structure and Regulation of Digital
Platforms (2019) .................................................... 26
vi
John Bergmayer, Even Under Kind Masters: A
Proposal to Require that Dominant Platforms
Accord Their Users Due Process (2018)................ 27
John Bergmayer, It Doesn’t Make Sense to Treat Ads
The Same As User Generated Content, Techdirt
(Aug. 17, 2020) ....................................................... 26
John Bergmayer, Section 230 Protects Speech, Not
Business Models, CPI Antitrust Chronicle (May
2021) ....................................................................... 19
Joseph Bernstein, Chaos on Twitter Leads a Group
of Journalists to Start an Alternative, NY Times
(Nov. 21, 2022),
https://www.nytimes.com/2022/11/21/style/mastodo
n-twitter-adam-davidson.html .............................. 24
Mastodon, Moderation actions,
https://docs.joinmastodon.org/admin/moderation 24
Mastodon, Servers, https://joinmastodon.org/servers
................................................................................ 25
Restatement (Second) of Torts (1977) ........................ 3
1
In the Supreme Court of the United States
___________
No. 21-1333
REYNALDO GONZALEZ, ET AL.,
PETITIONERS,
V.
GOOGLE LLC,
RESPONDENT.
___________
ON WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
___________
BRIEF OF PUBLIC KNOWLEDGE AS AMICUS CURIAE IN
SUPPORT OF RESPONDENT
___________
INTEREST OF AMICUS CURIAE1
Public Knowledge is a non-partisan, non-profit
consumer rights organization dedicated to promoting
freedom of expression, an open internet, and access to
affordable communications tools and creative works.
It has worked for many years to defend the procompetitive, pro-free expression goals of Section 230,
1 Pursuant to S. Ct. Rule 37.6, Public Knowledge states no
counsel for a party authored this brief in whole or in part and
no person or entity made a monetary contribution to its
preparation or submission.
2
while promoting policies designed to reduce online
harms.
SUMMARY OF ARGUMENT
1. Congress enacted Section 230 to permit
interactive computer services (“providers”) to exercise
editorial discretion when publishing third-party
content, without facing liability.
This case seeks to hold YouTube liable for
publishing objectionable third-party content. Section
230 does not allow this. Petitioners and some of their
amici try to work around this clear statutory
prohibition by characterizing their theory of liability
in different terms. They say that they seek to hold
YouTube liable, not for publishing third-party
content, but for “recommendations,” Pet. 29-33, or for
what the United States calls “YouTube’s own conduct
in designing and implementing recommendation
algorithms that result in the communication of a
distinct message from YouTube.” Brief for the United
States as Amicus Curiae in Support of Vacatur
(“United States”) 12. But this attempt to plead around
Section 230 must fail. YouTube’s conduct in this case
may be culpable, but that conduct was publishing and
immunized by Section 230. Characterizing content
recommendations as something other than
“publication,” or pleading causes of action that do not
name publication as an element does not change this.
Theories of liability that depend on the harmful
contents of third-party material constitute “treating”
a provider as a publisher and are barred by the
statute.
2. Section 230 protects the publication of thirdparty content, and it does so robustly. But it protects
only that. Just as some plaintiffs attempt to evade
Section 230 by characterizing their claims in other
terms, some defendant providers attempt to use
3
Section 230 as a defense in situations where it simply
does not apply. As this case is the Court’s first full
opportunity to consider the meaning and scope of
Section 230, it has the opportunity to provide clarity
to both lower courts and litigants as to both the reach,
and the limits, of Section 230.
3. Section 230 is intended to promote free
expression and competition online and to maximize
the ability of internet users to control the information
they see online. It is not a perfect statute, but its
fundamental policy goals of free expression,
competition, and user control are sound. Congress,
and not the courts, is the best avenue for policy
changes designed to better promote these goals while
reducing online harms.
ARGUMENT
I.
YouTube’s Recommendations Are a Form
of Publishing
Petitioners and their amici are correct to note that
Section 230 does not shield providers from liability for
their general business activities. Rather, it is better
understood to prevent liability for the act of
publishing in the common law sense. In the tort of
defamation, “publication” is an essential element.
Restatement (Second) of Torts § 558 (1977)
(“Restatement”). In this context, “publication” refers
to the dissemination or communication of material to
at least one other person (other than the person
defamed). The United States correctly explains that
publication “refers broadly to the communication of
expressive material to another.” United States 20
(citing Restatement § 577).
According to Petitioners,
4
YouTube selected the users to whom it would
recommend ISIS videos based on what YouTube
knew about each of the millions of YouTube
viewers, targeting users whose characteristics
suggested they would be interested in ISIS videos.
The selection of the users to whom ISIS videos
were recommended was determined by computer
algorithms created and implemented by YouTube.
Pet. 9 (citations omitted). This meets the common law
understanding of “publication” discussed above.
The fact that a publication is to a single individual,
uses an algorithm, or is targeted to specific users
based on a trove of personalized information does not
transform it into something other than a publication.
As the Second Circuit recently observed, the selection
of an audience, and matching content to likely
readers, has been a part of publishing content online
from the beginning, and “it would turn Section
230(c)(1) upside down to hold that Congress intended
that when publishers of third-party content become
especially adept at performing the functions of
publishers, they are no longer immunized from civil
liability.” Force v. Facebook, 934 F.3d 53, 67 (2d Cir.
2019). By recommending videos to users, YouTube is
disseminating them to new viewers. Communicating
expressive material to another is the essence of
publication, and presenting already-published
material to a new audience is considered a new
publication. See Giuffre v. Dershowitz, 410 F. Supp.
3d 564, 571 (S.D.N.Y. 2019) (“Republication to a new
audience or in a new forum does not come within the
single publication rule.”); Firth v. State, 98 N.Y.2d
365, 371 (2002) (“The justification for this exception
to the single publication rule is that the subsequent
publication is intended to and actually reaches a new
audience”); Salyer v. S. Poverty Law Ctr., 701 F. Supp.
2d 912, 916 (W.D. Ky. 2009) (same); cf. Banaian v.
5
Bascom, 281 A.3d 975, 980 (2022) (Section 230
protects Twitter users from claims of defamation for
republishing [retweeting], but not for their additional
commentary).
Of course, not all recommendations or references
to content (e.g., a positive movie review) constitute a
“publication” at all. As one court explained,
[T]he common thread of traditional republication
is that it presents the material, in its entirety,
before a new audience. A mere reference to a
previously published article does not do that.
While it may call the existence of the article to the
attention of a new audience, it does not present the
defamatory contents of the article to that
audience. Therefore, a reference, without more, is
not properly a republication.
Salyer, 701 F. Supp. 2d at 916. Unlike the references
discussed in Salyer, YouTube’s recommendation
systems present the contents of third-party content
directly to viewers. YouTube hosts the content itself,
and places recommended content before viewers in a
form where they can immediately view it. Because its
recommendations make the contents of the video
immediately available to viewers, they are
“publications” of user-uploaded videos under the
common law. Section 230 straightforwardly precludes
liability.
A.
Section 230 Precludes Liability
Under All Causes of Action That
Seek to Impose Liability for User
Material
Many forms of “recommendation” are not
republications. For example, “though a link and
reference may bring readers’ attention to the
existence of an article, they do not republish the
6
article.” In re Phila. Newspapers, 690 F.3d 161, 175
(3d Cir. 2012). But Section 230 precludes liability
under all causes of action that seek to impose liability
for recommendations, including those causes of action
that do not specifically name “publication” as an
element. By the terms of the statute, “No cause of
action may be brought and no liability may be imposed
under any State or local law that is inconsistent with
this section.” 47 U.S.C. 230(e)(3) (emphasis added).
This shows that Congress intended Section 230 to bar
any cause of action that seeks to hold a provider liable
for the contents of third-party material, however it is
phrased or re-characterized. Plaintiffs are creative
and can find causes of action that seek to treat
providers as “publishers” without expressly saying so.
In one case, “Plaintiffs do not contend that
[defendants] ‘published’ the tapes and pictures for
purposes of defamation and related theories of
liability... Instead, they say, [defendant] is liable for
‘negligent entrustment of a chattel.’” Doe v. GTE
Corp., 347 F.3d 655, 660 (7th Cir. 2003). But “a
plaintiff cannot sue someone for publishing thirdparty content simply by changing the name of the
theory from defamation to negligence.” Barnes v.
Yahoo!, 570 F.3d 1096, 1102 (9th Cir. 2009). “Nor can
he or she escape section 230(c) by [re]labeling . . . an
action that is quintessentially that of a publisher.” Id.
at 1102–03. Courts have interpreted Section 230 to
preclude word games of this sort and to give the
statute its intended effect. As the 9th Circuit found,
[W]hat matters is not the name of the cause of
action—defamation versus negligence versus
intentional infliction of emotional distress—what
matters is whether the cause of action inherently
requires the court to treat the defendant as the
“publisher or speaker” of content provided by
another. To put it another way, courts must ask
whether the duty that the plaintiff alleges the
7
defendant violated derives from the defendant’s
status or conduct as a “publisher or speaker.” If it
does, section 230(c)(1) precludes liability.
Id. at 1101–02. See also id. at 1101 (overview of cases
in different circuits where plaintiffs have advanced
various causes of action in attempts to work around
Section 230).
As the Fourth Circuit recently explained, treating
someone “as a publisher” means that liability is “on
account of some improper content within their
publication.” Henderson v. Source For Pub. Data, 53
F.4th 110, 122 (4th Cir. 2022). “[T]o hold someone
liable as a publisher at common law was to hold them
responsible for the content’s improper character.” Id.
Section 230’s prohibition on treating providers as the
publishers or speakers of third-party information
content therefore bars torts that require that the
provider have communicated or disseminated
objectionable material, or to have communicated or
disseminated material in a wrongful way. Any claim
where liability depends on the improper content of
third-party material is barred. See also Herrick v.
Grindr, 765 F. App’x 586, 590–91 (2d Cir. 2019)
(“manufacturing and design defect claims seek to hold
Grindr liable for its failure to combat or remove
offensive third-party content, and are barred by §
230”).
In this case, YouTube faces liability only because
of the content of some of the videos it recommended.
Plaintiff’s claims are precluded by Section 230,
because they inherently depend on the objectionable
contents of user-uploaded material, not just
YouTube’s own words or conduct. Thus, YouTube’s
content recommendations are shielded by Section 230
even if they are characterized as something other
than “publishing.”
8
Recommending content (either implicitly or
expressly) does not somehow adopt it such that it is
no longer “information provided by another
information content provider.” An “information
content provider” is “any person or entity that is
responsible, in whole or in part, for the creation or
development of information.” 47 U.S.C. 230(f)(3).
When a provider recommends third-party content, it
does not retroactively become its co-creator. As the
United States explains in its brief, the “development”
of information does not include measures providers
employ to make third-party information more
available to users without altering its content. United
States 21–23. In other words, the “development” of
material refers to its creation, not measures taken to
publish it to new audiences. Providers can always be
held responsible for material they have helped
create—for example, in Roommates, the provider was
liable because the unlawful content was a
“collaborative effort between Roommate and the
subscriber” at the time of its creation. See Fair Hous.
Council v. Roommates.com, 521 F.3d 1157, 1167 (9th
Cir. 2008) (“Roommates”). If the “development” of
material includes publishing, then section 230 would
be an empty statute.
Of course, if a provider so significantly transforms
third-party content so as to change its fundamental
character, it becomes new content created by the
provider. Batzel v. Smith, 333 F.3d 1018, 1031 (9th
Cir. 2003) (Section 230 applies when providers edit
third-party content as long as they do not change “its
basic form and message.”). Providers are always liable
for their own words, including those that refer to
third-party content. But a provider does not lose its
Section 230 protection for third-party content by
commenting on it, referencing it, or promoting it.
9
A provider that creates a “Recommended Videos of
the Week” list and writes its own descriptive copy
would be liable for its own words. But when a provider
recommends third-party videos, those videos remain
“information provided by another information content
provider.” Thus Petitioner’s insistence that “URLs
and notifications” are “information within the
meaning of Section 230(c(1),” Pet. 14, is immaterial.
Even if they are, they contain no objectionable content
in themselves. The objectionable content is the video,
not the URL for the video. “Information provided by
another information content provider” does not stop
being so when a provider gives it a new URL or
generates notifications to direct viewers to it. The
harms alleged in this case stem from the content of
ISIS videos. YouTube’s role was that of a publisher,
not a content creator. Therefore Section 230 precludes
Petitioner’s claims.
B.
Publishing Is Expressive Activity
That Often Conveys a Message of
Endorsement
Publishing is a typically expressive activity that
often
sends
a
message
of
agreement,
recommendation, or endorsement. While this is not a
First Amendment case, focusing on how publishing
can send a message in itself, in addition to (and
commenting on) the nature of what is being
published, can help clarify the nature of what Section
230 protects.
Analogizing print media to electronic media must
be done cautiously. “Accordingly, in search of cogent
principles, we compare the Internet to other media
with great care.” Oja v. United States Army Corps of
Engineers, 440 F.3d 1122, 1129 (9th Cir. 2006). But
while the scale and scope of internet publishing may
be relevant to policymakers, the same principles of
tort law apply. Thus, “the analogy between Internet
10
and print publication is sufficiently apt to be
serviceable.” Id. at 1131 (discussing defamation and
the single publication rule). In the print context, this
Court has already found that publication (including
what to publish, how to present it, and what not to
publish) are expressive activities protected by the
First Amendment. “The choice of material to go into a
newspaper, and the decisions made as to limitations
on the size and content of the paper, and treatment of
public issues and public officials—whether fair or
unfair—constitute the exercise of editorial control
and judgment.” Miami Herald Pub. Co., Div. of
Knight Newspapers. v. Tornillo, 418 U.S. 241, 258
(1974).
In this account by the Court, the action of
publication, the manner of presentation of material,
and the message being communicated are not each
distinct actions where liability could attach to the
presentation of material, but not its publication, or to
the message being expressed by the newspaper, but
not the means by which it expressed it. These are not
analytically distinct concepts and trying to
distinguish them leads to absurdities. It is illogical to
say that an entity cannot be held liable for publishing
objectionable material but can be held liable for the
manner in which it publishes, or for the “separate”
message the fact of its publication communicates. Yet
this is precisely the path urged by Petitioners and the
United States, who urge the Court to perceive a
distinction between “publishing” content and
“recommending” it. But these different concepts are
inextricable in practice.
The act of publication, without more, might
constitute a recommendation. When a newspaper
publishes a letter to the editor, it is communicating to
its readers that this letter, among the many it has
received, is worthy of attention. When it publishes an
11
op-ed—by placing it in the editorial section, copyediting it, and fitting it in with the overall design of
the newspaper—it is likewise sending a message of
endorsement (or even respectful disagreement), apart
from the content of the op-ed itself. The choice of what
to publish (and what not to publish), how to publish
it, and who to publish it to are inherently
communicative activities. Apart from First
Amendment concerns, holding a provider liable for
the expressive content of its publishing activities, but
not for the activities themselves, as the United States
suggests (at 27), is illogical.
According to the Second Circuit, attempting to
exclude “matchmaking” (a way of describing content
recommendations) from Section 230 would imperil
“the editorial decisions regarding third-party content
that interactive computer services have made since
the early days of the Internet.” Force, 934 F.3d at 66.
It continued,
The services have always decided, for example,
where on their sites (or other digital property)
particular third-party content should reside and to
whom it should be shown. Placing certain thirdparty content on a homepage, for example, tends
to recommend that content to users more than if it
were located elsewhere on a website. Internet
services have also long been able to target the
third-party content displayed to users based on,
among other things, users’ geolocation, language
of choice, and registration information. And, of
course, the services must also decide what type
and format of third-party content they will display
whether that be a chat forum for classic car lovers,
a platform for blogging, a feed of recent articles
from news sources frequently visited by the user,
a map or directory of local businesses, or a dating
service to find romantic partners.
12
Id. 66–67. This passage shows how Plaintiff and the
United States attempt to create a form of “publishing”
that does not include the selection of content for an
audience makes no sense. That simply is what
publishing consists of.
Any sense of what “publishing” means under
Section 230 includes full editorial discretion. The
Fourth Circuit recognized that Section 230(c)(1)’s
prohibition on treating providers as publishers means
that “lawsuits seeking to hold a service provider liable
for its exercise of a publisher’s traditional editorial
functions” with respect to user-submitted material
“such as deciding whether to publish, withdraw,
postpone or alter content—are barred.” Zeran v.
America Online, 129 F.3d 327, 330 (4th Cir. 1997).
(Section 230(c)(2) also covers content removals, and is
broader than (c)(1), because it immunizes content
removals from all causes of action, not just those that
“treat” a provider as a publisher or speaker. However,
unlike (c)(1), (c)(2) is subject to a “good faith”
limitation.)
This does not expand the protections of 230 beyond
“publishing” understood in the common law sense.
Consistent with the Supreme Court’s analysis in
Tornillo, Zeran is a recognition that “publishing” in
the narrowest sense is an expressive, communicative
act. And it recognizes that a liability shield for
publishing cannot just apply to the affirmative act of
publishing, but the failure to publish, the decision to
stop publishing, and the manner of publication—all
choices that inherently require subjective editorial
determinations, including whether to “recommend”
certain content through its manner of presentation.
YouTube’s content recommendations are protected
by Section 230 because they are the publication of
information provided by another information content
provider. Providers can be held liable for their own
13
conduct, but only to the extent that conduct is not
publishing, and liability does not depend on the
contents of third-party material. Internet publishers
like any others may intend to send a message of
endorsement or quality control in the selection of
what they publish and how they present it. But this
has no bearing on Section 230.
C.
Congress Enacted Section 230 To
Encourage Providers to Use Their
Independent Editorial Judgment to
Moderate Content
Section 230 was adopted to make it easier for
providers to refuse to publish material they find
objectionable, by removing “serious obstacles,” S. Rep.
No. 104-230, at 194 (1996), that would otherwise
disincentivize them from performing this editorial
function.
Section 230 was indisputably enacted to reduce
the amount of “obscene, lewd, lascivious, filthy,
excessively violent, harassing” content online, 47
U.S.C. § 230(c)(2), and to protect those providers who
“edit the smut from their systems.” Statement of Rep.
Goodlatte, 141 Cong Rec H 8460, 8471. It does this,
not only by stating that providers cannot be “treated”
as publishers of third-party content, but by
immunizing them from all torts (including those that
do not treat them as publishers) over removing
content they “consider[]… objectionable.” 47 U.S.C.
230(c)(2)(A). Both the words “consider” and
“objectionable” indicate that providers are expected to
use their own, independent editorial judgment in
determining both what criteria to apply, and how to
apply them. In addition to being a constitutional
necessity, ensuring that online publishers could
continue to use their own judgment to curate the
content on their services is the best way to achieve the
statute’s objective to protect the internet’s nature as
14
“a forum for a true diversity of political discourse,
unique opportunities for cultural development, and
myriad avenues for intellectual activity.” 47 U.S.C.
230(a)(3).
This is consistent with the legislative history. Rep.
Barton, for instance, saw the provision as offering
providers a way “to help them self-regulate
themselves without penalty of law.” Statement of Mr.
Barton, 141 Cong Rec H 8460, 8470. And numerous
statements in the legislative record saw Section 230
as a means for private actors, rather than the
government, to make editorial choices with respect to
online content. See Congressional Research Service,
Section 230: An Overview 5 (April 7, 2021).
Because Section 230 was intended to encourage
providers to take a more active role in curating thirdparty content, it should not be interpreted in a way
that penalizes providers who do so.
D.
“Recommendations” Are Not
Distinct from Other Features of
Interactive Computer Services
Petitioners and their amici advance a reading of
Section 230 that is so narrow that it would not apply
to nearly any feature of popular websites that existed
in 1996, much less today. They would effectively limit
Section 230 to services that acted as mere file hosts,
exercising little editorial control over what they
publish. But as the United States observes,
Interactive websites invariably provide tools that
enable users to create, and other users to find and
engage with, information. A chatroom might
supply topic headings to organize posts; a photosharing site might offer a feature for users to
signal that they like or dislike a post; a classifieds
website might enable users to add photos or maps
15
to their listings. If such features rendered the
website a co-developer of all users’ content, Section
230(c)(1) would be a dead letter.
United States 23. Section 230 was intended to apply
to interactive computer services. Therefore, as the
United States explains, the content discovery, search,
and organization features that interactive computer
services provide must be consistent with Section 230.
In other words, they are the publication, not the
“development,” of content.
This equally applies to the recommendation
features Petitioners, the United States, and others try
to distinguish. There is no basis in the statute to
conclude that an organizational tool that
“recommends” content should be treated any
differently. Such a tool could be, for example, a topic
heading of “Today’s Most Important Posts.” Or a
modern website might have “Today’s Most Important
Posts, Selected for You,” based on a user’s past
activity. It’s not that there is no clear line to be drawn
between organizing and “recommending” content; it’s
that there is no line to draw. To “recommend” content
is simply one way of organizing it.
Invocation of the word “algorithm” does not change
this. An algorithm is merely a sequence of steps that
a person or a computer might follow. In this context,
a provider’s algorithms determine how it organizes
content. One provider’s algorithm might be very
simple: “Display the most recent content first.”
Another’s might be more complex: “Display the most
clicked-on content first.” Section 230 protects a
provider’s publishing activities whether they are
performed by an algorithm it has deployed, or by
human moderators using their best judgment.
Of course, the “computer algorithms created and
implemented by YouTube,” Pet. 9, are enormously
16
more sophisticated than the ones mentioned above.
But they are still algorithms, different in degree, but
not kind, from simpler ones. Section 230 is a forwardlooking statute, which Congress expressly intended
“to promote the continued development of the
Internet.” 47 U.S.C. 230(b)(1). The Internet has
developed significantly since 1996, and there are good
public policy reasons to subject complex algorithms to
regulatory scrutiny. But that is not a matter for the
courts.
Petitioner’s attempt to distinguish YouTube’s
recommendations from search results is factually and
legally flawed. Pet. 15–16. Content recommendations
are not distinguishable from search results in a way
where Section 230 can be interpreted to cover one, but
not the other.
Search terms themselves might be algorithmically
auto-completed. Indeed, a “recommendation” may be
nothing more than a preemptively supplied search
result. Content recommendations, like providing
search results, are a way of communicating thirdparty material to a new audience. Section 230
immunizes providers from liability when they do this,
whether in search results or some other means.
Petitioner’s argument that Section 230 only
applies to actions “the user has actually requested,”
Pet. 15, does not withstand scrutiny. A platform
might publish material at the request of a user, or
unprompted. It is publishing in either case. Section
230 might only apply to “interactive computer
services,” but the behavior it shields, publishing, does
not have a requirement of interactivity. Indeed, the
suggestion that it does is squarely at odds with the
Petitioner’s claim (at 13, 32) that publishing should
be understood in its common law context. The
common law definition of publication does not hinge
17
on whether the recipients have requested the
material or not.
More broadly, search engines do not “only provide
users with materials in response to requests from the
users themselves.” Pet. 15. They attempt to provide
users with what they want, and different search
engines use different algorithms to both determine
and order (recommend) results. Search engines like
Google and Bing provide results customized to
individual users, based on their location, past search
history, and other criteria. By contrast, DuckDuckGo
argues that personalized results like this contribute
to “filter bubbles.” DuckDuckGo, Measuring the Filter
Bubble
(December
4,
2018),
https://spreadprivacy.com/google-filter-bubble-study.
Allowing competing search engines to develop their
own means to rank and order potential responses for
each individual is precisely the sort of “vibrant and
competitive free market,” 47 USC 230(b)(2), Congress
intended to promote via Section 230. It would make
no sense to impose liability for the kind of
differentiated,
individualized
determinations
Congress states it is the policy of Section 230 to
encourage.
II.
Section 230 Applies to Publishing, Not
Publisher Business Models
Section 230’s few words have profound effects.
People rely on the internet for news, entertainment,
and connecting with their friends and family. The
content moderation practices of social media sites
make national news. At the same time, section 230 is
not a deregulatory charter for the online businesses
that have become so central to culture, democracy,
and the economy. See Claire Cain Miller, When Uber
and Airbnb Meet the Real World, NY Times (Oct. 17,
2014), https://www.nytimes.com/2014/10/19/upshot/
when-uber-lyft-and-airbnb-meet-the-real-world.html.
18
Understanding what Section 230 covers also clarifies
what it does not.
A.
Section 230 Does Not Prevent
Holding Online Marketplaces
Accountable for Selling Dangerous
Products
Online marketplaces allow users to post items for
sale (or apartments to rent, or job vacancies to fill).
The different ways they are structured and
differences in state law mean that there is no simple
answer to the question of whether they are “sellers” of
goods transacted on their sites for the purposes of
consumer protection law and determining liability for
defective or dangerous products.
While it is true that these providers often function
as a forum for communication between buyers and
sellers, they also play a much more active role in
facilitating transactions. Online marketplaces often
handle payment processing, offer buyer protection
programs, and can even handle shipping and returns.
This level of involvement in the sale process goes
beyond simply publishing third-party content and
suggests that online marketplaces could be
considered a party to the sale.
This form of liability does not run afoul of Section
230 because holding someone liable as a seller is not
treating them as a publisher. See Oberdorf v.
Amazon.com Inc., 930 F.3d 136 (3rd Cir. 2019),
vacated and reh’g en banc granted, 936 F.3d 182 (3d
Cir. 2019); John Bergmayer, Section 230 Protects
Speech, Not Business Models, CPI Antitrust Chronicle
(May 2021). Selling goods on the internet is not
“publishing.” Moreover, seller liability does not
depend
on
the
contents
of
third-party
communications, even when the sale listing was
created by a user, and not the provider itself. Liability
19
would depend on the sale of a defective product, not
how it was described or who described it.
B.
Section 230 Does Not Prevent State
and Local Business Regulation
Section 230 does not prevent normal business
regulation, even of businesses that are in part
shielded by Section 230. While Section 230 prevents
state and local regulation of “publishing,” it does not
shield any company’s business model, and does not
prevent the imposition of liability on activities that
may be supported by or ancillary to “publishing,” but
are not publishing themselves.
For example, states and local governments are
free to regulate online transactions without
interference from Section 230. The Ninth Circuit
correctly pinpointed this distinction in a recent case
concerning local laws that imposed liability on shortterm home rental companies for booking unlawful
rentals that were initially posted by users. The court
noted “the Platforms face no liability for the content
of the bookings; rather, any liability arises only from
unlicensed bookings.” HomeAway.com v. City of Santa
Monica, 918 F.3d 676, 684 (9th Cir. 2018). Short-term
rental companies would be held liable for completing
transactions, not publishing third-party content.
Their liability does not depend necessarily on the
content of published material, and the publishing
components of their business can continue
unchanged.
C.
Section 230 Does Not Shield the
Collection and Use of Personal
Data
Section 230 shields providers from liability for
publishing third-party information content. As
discussed above, organizing, formatting, and
20
recommending content to users is “publishing” in the
common law sense. But Section 230 protects
publishing, not publisher business models, or
activities ancillary to publishing. For example, while
Section 230 shields publishers to the extent they are
targeting content to specific users, it does not stand in
the way of efforts to limit the collection of personal
information—even if a publisher plans to use that
information solely for publishing or collected it in the
course of publishing. The collection of data on
provider users is not the same thing as
communicating third-party content to an audience.
Nor does it authorize providers to use third-party
content or data derived from third-party content in
ways that go beyond publication or permit them to
escape liability for information content they
themselves have created or developed.
For example, in Brooks v. Thomson Reuters, the
defendant data broker argued that the creation of
“detailed cradle-to-grave dossiers” about “millions of
people” was shielded by Section 230. Brooks v.
Thomson Reuters Corp., No. 21-cv-01418-EMC, 2021
U.S. Dist. LEXIS 154093, at *2 (N.D. Cal. Aug. 16,
2021). The dossiers it sold contained “name[s],
photographs, personal identifying information, [and]
other personal data.” Id. The court found that these
dossiers were not “third-party content,” but “created
by Thomson Reuters, albeit from third-party sources.”
Id at *8, and that Section 230 did not apply. Id. at *39.
See also Roommates, 521 F.3d at 1171 (“Providing
immunity every time a website uses data initially
obtained from third parties would eviscerate the
exception to Section 230 for “develop[ing]” unlawful
content “in whole or in part.”) Section 230’s
protections apply to the extent that a provider is
publishing “information content,” actual expressive
material, from third parties. It does not permit them
to repackage data, facts, and claims from third parties
21
into their own expressive content and escape liability.
(For example, a YouTube user that repeats a
dangerous or defamatory claim in their video, in their
own words, cannot escape liability by later claiming
that the “information” in question was first “provided
by another information content provider.”)
III.
Section 230 Was Intended to Promote
Competition, Free Expression, and User
Control
In enacting Section 230, Congress recognized that
free, online communication would be stifled if
providers were responsible for each word their users
post and if the fear of legal liability prevented them
from moderating their services according to their
individual criteria. While the moderation choices of
online providers are protected by the First
Amendment, Section 230 makes it easier for providers
to exercise those rights. Section 230 also makes it
easier for Internet users to express themselves, and is
a pro-competitive statute intended to maximize their
sources of information online. However, a repeal of
the statute, or its substantial weakening either
legislatively or through judicial decree, would harm
both free expression and competition.
A.
Section 230 Protects and Promotes
a Competitive Market
Section 230 was intended “to encourage the
development of technologies which maximize user
control,” 47 U.S.C. 230(b)(3) over online information,
and to give online providers a freer hand in
moderating content. These goals work in tandem with
its pro-competitive purpose “to preserve the vibrant
and competitive free market that presently exists for
the Internet and other interactive computer services,”
47 U.S.C. 230(b)(2).
22
Without Section 230, smaller online providers
would be at a significant disadvantage due to their
limited resources. Any lawsuit brought against them
for the content of their users’ speech would have the
potential to bankrupt them, even if the suit was
ultimately found to be baseless. This risk of litigation
would make it much more difficult for smaller
providers to attract investment and grow their user
base. As discussed below, weakening Section 230
would make it more difficult for a new generation of
more decentralized, competitive online services to
emerge.
Larger online providers have the resources to
weather lawsuits and would be less affected by the
risk of litigation. Section 230 may have helped today’s
internet giants grow, but their continued dominance
has many causes, including lax antitrust enforcement
(see, e.g., Verizon Communs. v. Law Offices of Curtis
V. Trinko, 540 U.S. 398 (2004); Brooke Group v.
Brown & Williamson Tobacco, 509 U.S. 209 (1993)),
economies of scale (including economies of scale with
respect to content moderation), network effects, and
winner-take-all market dynamics. Careful public
policy is needed to bring about more online
competition. Making it riskier and more difficult for
new entrants to take on incumbents would have the
opposite effect.
The recent growth of the decentralized social
network Mastodon is an example of how Section 230’s
protections can foster a new generation of more
competitive, decentralized internet services. The
operator of each Mastodon “instance” sets its own
content moderation policy, which applies to posts
from its own users, as well as the content from other
instances that appears on their service. See
Mastodon,
Moderation
actions,
https://docs.joinmastodon.org/admin/moderation
23
(user guide for providers). An instance might have a
strict content moderation policy and might only
interoperate with instances that have similar policies.
Some instances are only open to specific kinds of users
(e.g., journalists). See Joseph Bernstein, Chaos on
Twitter Leads a Group of Journalists to Start an
Alternative,
NY
Times
(Nov.
21,
2022),
https://www.nytimes.com/2022/11/21/style/mastodontwitter-adam-davidson.html; Mastodon, Servers,
https://joinmastodon.org/servers (updated guide to
different instances). Other instances might have more
lax policies. Section 230 protects all these different
editorial choices, and the ability of users to select
providers with policies they agree with. The sole
operator of a Mastodon instance could not take on
liability as the “publisher” of the thousands of pieces
of content that might flow across her service daily. Yet
services such as Mastodon were not only what
Congress squarely intended to promote in 1996, but a
more decentralized approach to social media might be
an important solution to combating the abuses and
deleterious social role of current major platforms.
Such an approach would be imperiled if Plaintiffs
prevail in their claims.
B.
Only Congress Can Update Section
230 to Better Fulfill Its Purposes in
Today’s Circumstances
Different Courts of Appeal have independently
interpreted Section 230 in a way that Petitioners
challenge, not because they have all fallen prey to the
same interpretive error, but because they are
compelled by the logic and words of the statutory text,
along with the legislative history behind the statute.
The job of the courts is to give effect to Congress’
choices, not to substitute their own.
Section 230 like any other law should be
periodically reexamined and updated when
24
circumstances change. Possible revisions can help it
address new online challenges and better achieve its
pro-competitive, pro-free expression goals. More than
25 years later, it may be wise to revisit these choices.
But that is the job of Congress, not the judiciary. See
Arizona v. Mayorkas, 598 U.S. __ (2022), Gorsuch, J.,
statement respecting denial of certiorari (“We are a
court of law, not policymakers of last resort.”). The
Court is not made up of legislative draftsmen nor
legislative custodians, cleaning up the messes and
mistakes of legislators past. Any fixes to Section 230
must come from Congress. Indeed, there is a litany of
proposed Section 230 amendments in both the past
and current Congress, see, e.g., Platform Integrity Act,
H.R. 9695, 118th Cong. (2022); PACT Act, S. 797,
117th Cong. (2021), and Congress recently enacted
the Allow States and Victims to Fight Online Sex
Trafficking Act of 2017, PL. 115-164 (2017).
Present amicus has supported regulating
platforms to better achieve Section 230’s goals. See
John Bergmayer, It Doesn’t Make Sense to Treat Ads
The Same As User Generated Content, Techdirt (Aug.
17, 2020), https://www.techdirt.com/2020/08/17/itdoesnt-make-sense-to-treat-ads-same-as-usergenerated-content (proposing a legislative change to
Section 230). As discussed above, reinvigorated
antitrust policy will help check the market power of
dominant providers, allowing users more choice. A
new sector-specific digital regulator would have the
flexibility to address a tech industry that is complex
and subject to different competitive chokepoints, see
Harold Feld, The Case for the Digital Platform Act:
Market Structure and Regulation of Digital Platforms
(2019), https://publicknowledge.org/policy/the-casefor-the-digital-platform-act/, and granting users of
dominant providers due process rights can promote
the goals of free expression. See John Bergmayer,
Even Under Kind Masters: A Proposal to Require that
25
Dominant Platforms Accord Their Users Due Process
(2018), https://publicknowledge.org/policy/paper-ondominant-online-platforms-and-due-process.
As this case shows, the availability and reach of
harmful content online is a serious problem.
Providers do not always use the flexibility afforded to
them by the law wisely. In this case, material that
violates YouTube’s terms of service was instead
promoted and recommended to users. That content
promoted and glorified terrorism. Other cases have
involved providers endangering the lives of children
through easy access to pro-suicide content and even
the means to efficiently commit suicide. See, e.g.,
Dyroff v. Ultimate Software Grp., Inc., 934 F.3d 1093
(9th Cir. 2019); Anderson v. TikTok, 2022 U.S. Dist.
LEXIS 193841 (E.D. Pa. 2022); Maynard v. Snapchat,
357 Ga. App. 496 (Ga. Ct. App. 2020). These are grave
societal problems that demand attention from
policymakers.
However, addressing these issues this must be a
job for Congress, not the courts. There is no way to
creatively interpret Section 230 to cover some forms
of publishing but not others, or to allow liability for
third-party content in some cases but not others, or
for some kinds of digital providers, but not others.
Any judicially-created loopholes would inevitably
swallow the statute, as plaintiffs and lower courts
would have little trouble expanding any loophole to
cover new circumstances. The theory put forth by
Petitioners in this case and supported by various
amici, whereby a provider becomes liable for content
if it “recommends” it, is one such theory. It does not
require much imagination to see how nearly any
means YouTube selects to present videos to users
could be framed as a “recommendation”—for example,
any video on a user’s front page.
26
More than that, under Petitioner’s theory, users
(and potential plaintiffs) would not know, on seeing a
video, whether YouTube could be held liable or not,
since YouTube might be potentially liable if it shows
a user a video one way, but not another way, or liable
to one set of users but not another. At best, judicially
revising Section 230 to exclude recommendations
would cause chaotic results. Congress and not the
courts can best craft policies that are not only sound
from a policy perspective, but administrable,
narrowly targeted, and that protect free expression.
CONCLUSION
For the above reasons, the Court should reject
Petitioner’s claims and uphold the court below.
Respectfully submitted,
JOHN BERGMAYER
ALEX PETROS
HAROLD FELD
Counsel of Record
PUBLIC KNOWLEDGE
1818 N St. NW Suite 410
Washington, DC 20782
(202) 861-0020
hfeld@publicknowledge.org
Counsel for Amicus Curiae
This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.