Amicus Curiae Brief — Cox Communications, Inc., et al., Petitioners v. Sony Music Entertainment, et al.

Supreme Court briefSep 5, 2025

Ask Donna

What actually matters in this document.

Text

No. 24-171

In the

Supreme Court of the United States

COX COMMUNICATIONS, INC., et al.,

Petitioners,

SONY MUSIC ENTERTAINMENT, et al.,

Respondents.

On Writ of Certiorari to the United States

Court of A ppeals for the Fourth Circuit

BRIEF AMICUS CURIAE OF

FLOOR64, INC. D/B/A THE COPIA INSTITUTE

IN SUPPORT OF PETITIONERS

Catherine R. Gellis, Esq.

Counsel of Record

3020 Bridgeway, #247

Sausalito, CA 94965

(202) 642-2849

cathy@cgcounsel.com

Counsel for Amicus Curiae

120598

(800) 274-3321 • (800) 359-6859

TABLE OF CONTENTS

Page

TABLE OF CONTENTS

INTEREST OF AMICUS CURIAE . . . . . . . . . . . . . . . 1

SUMMARY OF ARGUMENT . . . . . . . . . . . . . . . . . . . . 3

ARGUMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

A legal regime that could hold Cox liable

offends the First Amendment . . . . . . . . . . . . . . . . 4

A. There is a prior restraint problem . . . . . . . . 5

B. There is a Packingham problem . . . . . . . . 14

C. There is a Vullo problem . . . . . . . . . . . . . . . 18

II. A legal regime that could hold Cox liable

would affect all online expression . . . . . . . . . . . 20

A. Targeting platforms with secondary

liability inherently affects user expression . . 20

B. There is nothing about copyright,

filesharing, or Cox that justifies

secondary liability . . . . . . . . . . . . . . . . . . . . 26

CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

TABLE OF CITED AUTHORITIES

Page

CASES:

Bantam Books, Inc. v. Sullivan,

372 U.S. 58 (1963) . . . . . . . . . . . . . . . . . . . . . . . 5, 13, 14

BMG Rts. Mgmt. (US) LLC v.

Cox Commc’ns, Inc.,

149 F. Supp. 3d 634 (E.D. Va. 2015) . . . . . . . . . . . . . . 8

BMG Rts. Mgmt. (US) LLC v.

Cox Commc’ns, Inc.,

881 F.3d 293 (4th Cir. 2018) . . . . . . . . . . . . . . . . 7, 8, 11

Concord Music Group, Inc. v. X Corp.,

No. 3:23-cv-00606 (M.D. Tenn. Mar. 5, 2024) . . . . . 16

Eldred v. Ashcroft,

537 U.S. 186 (2003) . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Fair Housing Coun. Of San Fernando Valley v.

Roommates. Com, LLC,

521 F.3d 1157 (9th Cir. 2008) . . . . . . . . . . . . . . . . . . . 23

Gonzalez v. Google,

598 U.S. 617 (2023) . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Lenz v. Universal Music Corp.,

815 F.3d 1145 (9th Cir. 2015), cert. denied,

137 S. Ct. 2263 (2017) . . . . . . . . . . . . . . . . . . . . . . . . . 10

iii

Cited Authorities

Page

Metro-Goldwyn-Mayer Studios Inc. v.

Grokster, Ltd.,

545 U.S. 913 (2005) . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Moody v. NetChoice,

144 S. Ct. 2383 (2024) . . . . . . . . . . . . . . . . . . . . 4, 12, 19

Murthy v. Missouri,

144 S. Ct. 1972 (2024) . . . . . . . . . . . . . . . . . . . . . . . . . 19

Nat. Rifle Association of America v. Vullo,

144 S. Ct. 1316 (2024) . . . . . . . . . . . . . . 13-14, 18, 19, 30

Near v. Minnesota ex rel. Olson,

283 U.S. 697 (1931) . . . . . . . . . . . . . . . . . . . . 7, 11-13, 19

Packingham v. North Carolina,

137 S. Ct. 1730 (2017) . . . . . . . . . . . . . . . . . . . . . . . 14-17

Perfect 10, Inc. v. CCBill LLC,

488 F.3d 1102 (9th Cir. 2007) . . . . . . . . . . . . . . . . . . . . 7

Sony Corp. of America v. Universal City

Studios, Inc.,

464 U.S. 417 (1984) . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Twitter, Inc. v. Taamneh,

143 S. Ct. 1206 (2023) . . . . . . . . . . . . . . . . . . . . . . . . . 28

UMG Recordings, Inc. v. Shelter Capital Partners,

718 F.3d 1006 (9th Cir. 2013) . . . . . . . . . . . . . 22, 23, 25

Cited Authorities

Page

Zeran v. America Online, Inc.,

129 F.3d 327 (4th Cir. 1997) . . . . . . . . . . . . . . 22, 23, 26

STATUTES AND OTHER AUTHORITIES:

U.S. Const., amend. I . . . . . . . . . . . . 1-5, 15, 18, 19, 24, 27

U.S. Const., art. I, § 8, cl. 8 . . . . . . . . . . . . . . . . . . . . . . . 15

17 U.S.C. § 506 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

17 U.S.C. § 512(a) . . . . . . . . . . . . . . . . . . . . . . . . . . 6, 12, 20

17 U.S.C. § 512(c) . . . . . . . . . . . . . . . . . . . . . .6, 8, 12, 16, 19

17 U.S.C. § 512(c)(1)(C) . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

17 U.S.C. § 512(c)(3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

17 U.S.C. § 512(c)(3)(A)(vi) . . . . . . . . . . . . . . . . . . . . . . . . 10

17 U.S.C. § 512(d) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

17 U.S.C. § 512(d)(3) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

17 U.S.C. § 512(f) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

17 U.S.C. § 512(g)(2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

17 U.S.C. § 512(i)(1)(A) . . . . . . . . . . . . . . . . . . . . . . . . . . 6-7

Cited Authorities

Page

17 U.S.C. § 512(m)(1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

47 U.S.C. § 230 . . . . . . . . . . . . . . . . . . . . . . . . 21, 23, 26-28

47 U.S.C. § 230(c) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

47 U.S.C. § 230(c)(1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

47 U.S.C. § 230(e)(2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

A Note About the Knight Institute, X, and Bluesky,

Knight First A mendment Institute (Nov. 25,

2024), https://knightcolumbia.org/blog/a-noteabout-the-knight-institute-x-and-bluesky . . . . . . . . . 1

Karl Bode, Two Decades Later And The FCC

Is Still Trying To Crack Down On AntiCompetitive Deals Between Landlords And

Broadband Monopolies, Techdirt (Mar. 18,

2024), https://www.techdirt.com/2024/03/18/

two-decades-later-and-the-fcc-is-still-tryingto-crack-down-on-anti-competitive-dealsbetween-landlords-and-broadband-monopolies/ . . . 17

Brief amicus curiae of M. Chris Riley, Floor64, Inc.

d/b/a The Copia Institute, and Engine Advocacy,

Gonzalez v. Google, 598 U.S. 617 (2023) (21-1333) . . 20

Cited Authorities

Page

Engine, Startups, Content Moderation, & Section

230, available at https://static1.squarespace.

com/static/571681753c44d835a440c8b5/t/6

1b26e51cdb21375a31d312f/1639083602320/

S t a r t u p s % 2 C + C o n t e n t+M o d e r a t i o n

%2C+and+Section+230+2021.pdf . . . . . . . . . . . . . . 22

Timothy Geigner, Bungie, YouTuber Settle

Lawsuit Over Fraudulent DMCA Takedowns

For YouTube Videos, Techdirt (Mar. 29, 2024),

https://www.techdirt.com/2024/03/29/bungieyoutuber-settle-lawsuit-over-fraudulentdmca-takedowns-for-youtube-videos/ . . . . . . . . . . . . . 9

Cathy Gellis, Creating State Action Via Antitrust

Law An d Ma kin g Th e Peo pl e W h o’ve

Been Wrong About The Constitutionality

Of Content Moderation Suddenly Right,

T ec h di rt (Jun. 2 8 , 2 021), https://w w w.

techdir t.com /2021/0 6/2 8/creating-stateaction-via-antitr ust-law-making-peoplewhove-been-wrong-about-constitutionalitycontent-moderation-suddenly-right/ . . . . . . . . . . . . . 25

Cathy Gellis, Senator Klobuchar Proposes An

Unconstitutional Law That Would Kill

Legions Of People If Tr ump Were Still

President, Techdirt (Jul. 26, 2021), https://www.

techdirt.com/2021/07/26/senator-klobucharproposes-unconstitutional-law-that-would-killlegions-people-if-trump-were-still-president/ . . 24-25

vii

Cited Authorities

Page

Eric Goldman, In 512(f), the “F” Stands for

“Fu ti li ty”, T e c h n o l o g y & M a r k e t i n g

L aw B l o g (Aug. 15, 2025), https:// blog.

ericgoldman.org/archives/2025/08/in-512f-thef-stands-for-futility-shaffer-v-kavarnos.htm . . . . . . 11

Peter Guest, Exposed documents reveal how the

powerful clean up their digital past using a

reputation laundering firm, Rest of World

(Feb. 3, 2022), https://restofworld.org/2022/

documents-reputation-launder ing-f i r meliminalia/ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27-28

Kashmir Hill, An Internet Veteran’s Guide

to Not Being Scared of Technology, New

York T imes (Jul. 29, 2023), https://w w w.

nytimes.com/2023/07/29/technology/mike

masnick-techdirt-internet-future.html) . . . . . . . . . . . 1

Stephanie Kirchgaessner, Top FDA official

demands removal of YouTube videos in which he

criticized Covid vaccines, The Guardian (Aug.

31, 2025), https://www.theguardian.com/usnews/2025/aug/31/fda-official-youtube-videos . . . 9, 10

Mike Masnick, Amy Klobuchar Wants To Break

The Internet Because Someone Made A

Stupid Satirical Video About Her, Techdirt

(Aug. 21, 2 0 2 5), https: // w w w.t echd i r t .

com/2025/08/21/amy-klobuchar-wants-tobr e a k- t he - i nt e r n e t - b e c a u s e - s o m e on e made-a-stupid-satirical-video-about-her/ . . . . . . . . 24

viii

Cited Authorities

Page

Mike Masnick, Announcing Ctrl-Alt-Speech:

A New Pod ca st Abo ut O nlin e Speech,

T e c h di r t (Ma r. 7, 2 0 2 4), https: // w w w.

techdirt.com/2024/03/07/announcing-ctrl-altspeech-a-new-podcast-about-online-speech/ . . . . . . . 1

Mike Masnick, Bipartisan Group Of Senators

Introduce New Terrible ‘Protect The Kids

O nlin e’ Bill, T e c h di r t (May 2 , 2 0 2 4),

ht t p s: // w w w.t e c hd i r t .c om / 2 0 2 4 / 0 5 / 0 2 /

bipartisan-g roup-of-senators-introducenew-terrible-protect-the-kids-online-bill/ . . . . . . . . 24

Mike Masnick, Just A Click Away: How To Improve

Broadband Access, Techdirt (Jul. 18, 2022), https://

www.techdirt.com/2022/07/18/just-a-click-awayhow-to-improve-broadband-competition/ . . . . . . . 16-17

Mike Masnick, Protocols, Not Platforms: A

Technological Approach to Free Speech,

Techdirt (Aug. 28, 2019), https://www.techdirt.

com /2019/08/28/protocols-not-platformstechnological-approach-to-free-speech/ . . . . . . . . . . . 1

Mike Masnick, Success! One Billion Users Will

Go Into Production (Late Backers Welcome),

Techdirt (Dec. 20, 2024), https://www.techdirt.

com/2024/12/20/success-one-billion-userswill-go-into-production-late-backers-welcome/ . . . . . 2

Cited Authorities

Page

Mike Masnick, The Sky Is Rising 2024 Edition:

Rather Than Destroying Culture , The

Internet Has Saved The Content Industries,

Techdirt (Jan. 22, 2024) https://www.techdirt.

com / 2 0 2 4 /01 / 2 2 /the - sk y-is-r isi ng-2 0 2 4 edition-rather-than-destroying-culture-theinternet-has-saved-the-content-industries/ . . . . . 15-16

Mike Masnick, Trump Promises To Abuse Take It

Down Act For Censorship, Just As We Warned,

Techdirt (Mar. 6, 2025), https://www.techdirt.

com/2025/03/06/trump-promises-to-abuse-takeit-down-act-for-censorship-just-as-we-warned/ . . . 24

Mike Masnick, Trust & Safety Professional

Association Launches: This Is Important,

Techdirt (Jun. 17, 2020), at https://www.techdirt.

com/2020/06/17/trust-safety-professionalassociation-launches-this-is-important/ . . . . . . . . . . 25

Mike Masnick, Trust & Safety Tycoon: Try

Your Hand At Managing A Social Media

Trust & Safety Team, T echdirt (Oct. 17,

2023), https://www.techdirt.com/2023/10/17/

t r u st - s a fet y- t yc o on - t r y-you r - h a nd - at managing-a-social-media-trust-safety-team/ . . . . . . 2

Cited Authorities

Page

Mike Masnick, Veoh Still Not Dead Enough

For Universal Music; Asks Court To Rehear

Case Yet Again, T echdirt (Feb. 11, 2010),

https://www.techdirt.com/2013/04/08/veohst i l l-not - dead- enough-u n iversa l-music asks-court-to-rehear-case-yet-again/ . . . . . . . . . . . 22

Andy Maxwell, TV Museum Will Die in 48

Ho urs Unl ess So ny Retra cts Yo uTube

2023), https://torrentfreak.com/tv-museumwill-die-in-48-hours-unless-sony-retractsyoutube-copyright-strikes-230904/ . . . . . . . . . . . . . 10

Toby Shulruff, Jeff Lazarus, and A manda

Menking, Voices of Trust & Safety: Origins

and Evolutions (2025), https://papers.ssrn.

com/sol3/papers.cfm?abstract_id=5361961 . . . . . . . 25

Jennifer M. Urban, Joe Karaganis, & Brianna

L. Schof ield, Notice and Takedow n in

Everyday Practice (UC Berkeley Public

L a w R e s e a r c h P a p e r No . 2 7 5 5 6 2 8 ,

2017), https://ssrn.com/abstract=2755628 . . . . 8-10, 12

INTEREST OF AMICUS CURIAE1

Amicus Copia Institute submits this brief to highlight

the ways this dispute implicates the First Amendment

by affecting online speech and the ability for anyone to

facilitate it.

The Copia Institute is the think tank arm of Floor64,

Inc., the privately-held California small business behind

Techdirt.com (“Techdirt”). 2 As a think tank the Copia

Institute produces evidence-driven articles and papers3 as

well as other forms of expressive output such as podcasts4

1. No counsel for any party authored this brief in whole or

in part. Amicus and its counsel authored this brief in its entirety.

No person or entity other than amicus and its counsel made

a monetary contribution intended to fund the preparation or

submission of this brief.

2. See Kashmir Hill, An Internet Veteran’s Guide to Not

Being Scared of Technology, New York T imes (Jul. 29, 2023),

https://www.nytimes.com/2023/07/29/technology/mike masnicktechdirt-internet-future.html (profiling its founder, owner, and

editor Michael Masnick).

3. See, e.g., Mike Masnick, Protocols, Not Platforms: A

Technological Approach to Free Speech, Techdirt (Aug. 28, 2019),

https://www.techdirt.com/2019/08/28/protocols-not-platformstechnological-approach-to-free-speech/. This paper is credited

with inspiring the development of the Bluesky social media

platform. A Note About the Knight Institute, X, and Bluesky,

K night First A mendment I nstitute (Nov. 25, 2024), https://

knightcolumbia.org/blog/a-note-about-the-knight-institute-xand-bluesky.

4. See, e.g., Mike Masnick, Announcing Ctrl-Alt-Speech:

A New Podcast About Online Speech, Techdirt (Mar. 7, 2024),

https://www.techdirt.com/2024/03/07/announcing-ctrl-alt-speecha-new-podcast-about-online-speech/.

and games5 that examine the nuances and assumptions

underpinning technology policy. Armed with its insights

it then regularly submits advocacy instruments such as

regulatory comments and amicus briefs, including before

this Court, with recent submissions in Tiktok v. Garland,

No. 24-696, Murthy v. Missouri, No. 23-411, Moody v.

NetChoice, No. 22-277, and Gonzalez v. Google, No. 211333, which all involved similar issues as those present

here: the expressive rights of Internet platforms and those

that use them.6

The goal of all of the Copia Institute’s efforts is to

educate courts, lawmakers, and other regulators—as

well as innovators, entrepreneurs, and the public—and

influence good policy that promotes and sustains innovation

and expression. It also does so through Techdirt, an online

5. See, e.g., Mike Masnick, Trust & Safety Tycoon: Try

Your Hand At Managing A Social Media Trust & Safety Team,

T echdirt (Oct. 17, 2023), https://www.techdirt.com/2023/10/17/

trust-safety-tycoon-tr y-your-hand-at-manag ing-a-socialmedia-trust-safety-team/ (adding “Trust and Safety Tycoon” to

“Moderator Mayhem” in its game catalog, each enabling players

to experience the difficulties of effective platform moderation

given various competing pressures that typically bear on the site

management experience). It also recently funded the development

of its latest game, “One Billion Users,” thanks to the use of other

Internet platforms including Kickstarter and Bluesky. Mike

Masnick, Success! One Billion Users Will Go Into Production

(Late Backers Welcome), Techdirt (Dec. 20, 2024), https://www.

techdirt.com/2024/12/20/success-one-billion-users-will-go-intoproduction-late-backers-welcome/.

6. The Copia Institute also submitted a brief in Andy Warhol

Foundation Visual Arts v. Goldsmith, No. 21-869, discussing the

First Amendment issues involved in copyright.

publication that has chronicled technology law and policy

for nearly 30 years. In this time it has published more

than 80,000 articles regarding subjects such as freedom

of expression, copyright, and platform moderation—issues

at the heart of this matter—as well as other related topics

including cybersecurity, competition policy, and the impact

of technology on civil liberties generally. Furthermore,

the company is not just a speaker regularly availing itself

of its First Amendment protections to engage in its own

expression but also a platform provider itself, soliciting

what has amounted to over two million reader comments,

which is a form of user expression that advances discovery

and discussion. The company then uses other Internet

platforms of various types to promote its own expression

and engage with its audiences.

The Copia Institute therefore submits this brief

wearing two hats: as a longtime commenter on the issues

at the heart of this challenge, and as an example of those

whose own First Amendment rights, as both a platform

operator and user, are threatened by a legal regime that

could hold Petitioner Cox Communications (“Cox”) liable,

as the lower courts have allowed here.

SUMMARY OF ARGUMENT

Amicus Copia Institute submits this brief amicus

curiae to caution that this case not be treated as just a

Millennium Copyright Act (“DMCA”) case. The dispute

here is not so esoteric; rather, this is a case whose outcome

will ultimately bear on all online expression in both its

practical and legal effects because, at its heart, it is a case

about secondary liability for the online platforms7 that

facilitate online expression—including, ultimately, any

form of secondary liability arising from their users—and

whether that secondary liability can constitutionally be a

basis for regulatory pressure causing platforms to police

user expression. The answer must be no, as subjecting

them to such liability regimes, as Cox was here, abrogates

both platforms’ own First Amendment rights and those

of their users.

ARGUMENT

A legal regime that could hold Cox liable offends

the First Amendment

Answering the question of whether Cox should

be liable for its users’ activity requires asking what,

if anything, Cox should have done differently to avoid

liability, and what the consequences would have been

to its speech interests, and those of its users, if it had.

Because if the only way to avoid liability is for a platform

to act against those interests, effectively under penalty

of law and significant liability if it does not, then that law

is unconstitutional.

7. The lower courts have tended to use “service providers,”

or just “providers” to describe the intermediaries who enable

expression to be shared and exchanged on the Internet because

this is the term used by the DMCA, the statute Cox had sought

to invoke at an earlier stage of this dispute to insulate it from

charges of secondary liability. This brief, however, frequently uses

“platform” interchangeably with “provider,” as has this Court in

much of its relevant jurisprudence related to online speech issues,

to drive home the point that the issues raised here are exactly the

same as those this Court has addressed in these other cases. See,

e.g., Moody v. NetChoice, 144 S. Ct. 2383, 2393 (2024).

There is a prior restraint problem

This Court has long held it anathema to the First

Amendment for speakers and speech to suffer punitive

consequences—a prior restraint—before there has

been proper adjudication to determine whether those

consequences are deserved. See Bantam Books, Inc. v.

Sullivan, 372 U.S. 58, 70 (1963). Yet, like in Bantam Books

v. Sullivan, this case is all about punitive consequences

to speech interests flowing from what were but mere

accusations and not “judicial determinations” of actual

wrongdoing. Id. (holding it was prior restraint to subject

expression to the judgment of non-judicial decisionmakers).

The case against Cox is premised on the idea that, when

presented with these accusations of its users’ alleged

wrongdoing, it did not itself inflict a punitive consequence

upon them, in this case by not terminating their accounts,

and, because it did not, it was then due its own punitive

consequences, first losing its safe harbor protection for

not having exercised this extrajudicial enforcement on its

users, and then finding itself in line for shared liability on

a legal theory premised largely on its awareness of those

accusations.

But the fundamental problem with accusations is

that they are often wrong, which is why due process is

required to determine whether an accusation is sound

before a resulting legal consequence can be meted out.

See Bantam, 372 U.S. at 71-72. So when the only way for

Cox to have avoided this trouble is if it had automatically

treated every accusation as valid, with no adjudication, it

is problematic on several fronts.

One is as a matter of statutory construction, because

allowing Cox to be liable at all misconstrues the language

of the DMCA in at least two ways. One way is that Cox, as

a provider of “transitory digital network communications,”

was entitled to the 512(a) safe harbor. 17 U.S.C. § 512(a). But

this safe harbor is not conditioned on the provider doing

anything in response to accusations of infringement. Id. In

fact, this safe harbor imposes no obligation to even receive

them at all. See id. The “takedown notices” that the DMCA

authorizes to convey these infringement accusations to

platforms are entirely creatures of the 512(c) safe harbor,

which is used by platforms that store information “at the

direction of users.” 17 U.S.C. § 512(c)(3). 8 As a result, even

if every takedown notice Cox ever received had been a

valid accusation that it ignored, its safe harbor against

any claim of liability should never have been in jeopardy.9

The other chief way that Cox’s fate was inconsistent

with the language of the DMCA is that, even for providers

using those safe harbors conditioned on responding to

takedown notices, none are conditioned on terminating

users in response to infringement allegations. The statute

only requires there to be a policy for terminating repeat

infringers “in appropriate circumstances.” 17 U.S.C.

8. The 512(d) safe harbor for providers of information location

tools is also partially conditioned on receiving 512(c) takedown

notices. 17 U.S.C. § 512(d)(3). But the 512(a) safe harbor makes

no such cross-reference.

9. It also means that any accusation sent to Cox via a

“takedown notice” was simply just an accusation, rather than any

sort of formal notice with statutory force. But because providers

using 512(c) and (d) safe harbors do receive accusations via formal

notices, it is important to remain focused on the fact that, formal

or not, these notices are still just accusations of wrongdoing and

not adjudicated findings.

§ 512(i)(1)(A) (“The limitations on liability established by

this section shall apply to a service provider only if [it]

has adopted and reasonably implemented […] a policy that

provides for the termination in appropriate circumstances

of subscribers and account holders […] who are repeat

infringers.”). Courts had previously read this provision to

afford the platforms flexibility in how they responded to

repeat infringers. See Perfect 10, Inc. v. CCBill LLC, 488

F.3d 1102, 1109-10 (9th Cir. 2007) (“The statute permits

service providers to implement a variety of procedures,

but an implementation is reasonable if, under ‘appropriate

circumstances,’ the service provider terminates users who

repeatedly or blatantly infringe copyright.”). But that

language about “repeat infringers” must be interpreted

as adjudicated infringers. Otherwise it would require

providers to inflict an unconstitutional prior restraint on

their users if those users could be deemed infringers based

solely on accusation, and effectively punished for it by being

terminated, without a court ever having weighed in.10

Cox’s experience also illustrates why interpreting the

provision without that implicit “adjudicated” caveat would

be so constitutionally problematic. During the period

that the Respondent was sending its takedown notices,

others were sending theirs. BMG Rts. Mgmt. (US) LLC

v. Cox Commc’ns, Inc., 881 F.3d 293, 299-300 (4th Cir.

2018). Cox’s alleged failure to act on them—in particular,

by not terminating users targeted by them—became the

10. Termination also functions as a prospective punishment,

against future infractions, and is thus an unconstitutional prior

restraint for that reason as well. Near v. Minnesota ex rel. Olson,

283 U.S. 697, 720 (1931) (“Subsequent punishment for such abuses

as may exist is the appropriate remedy[.]”).

basis of an earlier lawsuit, BMG Rights Management

v. Cox Communications, which resulted in Cox losing

its safe harbor protection. Id. at 300. But that litigation

itself revealed that many of the takedown notices Cox

had received were in fact invalid—in fact, they were so

invalid that an entire set of plaintiff claims were dismissed

from the lawsuit because the plaintiff had no right to

enforce a copyright in the implicated works. BMG Rts.

Mgmt. (US) LLC v. Cox Commc’ns, Inc., 149 F. Supp. 3d

634, 643-653 (E.D. Va. 2015). Nevertheless, despite the

presence of so many invalid claims, it was held against

Cox, in BMG as well as this case, that it had not treated

all the other infringement accusations it had received as

presumptively valid, even though none of them had ever

been legally tested and, given the invalidity of so many,

Cox had reason to be skeptical of their legitimacy. See, e.g.,

C.A.App.538 (“[W]e have seen issues of false allegations

against subscribers. [W]e have documented cases where

complaints came in against IP addresses that weren’t even

in use in our network. So we know that some portion of

the complaints that we get are false accusations.”).

But amicus Copia Institute writes because it is not

just Cox that can find itself at the receiving end of invalid

takedown notices. The danger of a regulatory scheme that

requires platforms to presume all accusations as valid is

that they are so often not. See, e.g., Jennifer M. Urban, Joe

Karaganis, & Brianna L. Schofield, Notice and Takedown

in Everyday Practice (“Urban”) 88 (UC Berkeley Public

Law Research Paper No. 2755628, 2017), https://ssrn.com/

abstract=2755628. And not just in this case, with respect

to this provider, or this sort of conduit provider, but in

all cases, with all providers, including those using the

512(c) safe harbor as platforms hosting stored expression,

where these takedown notices conveying mere accusations

result in the removal of that speech, if not also eventually

the speaker behind it.11 These platforms help weave the

rich fabric of the Internet, offering any sort of Internet

service that can be imagined, including social media,

video hosting, music hosting, photo hosting, general cloud

storage, consumer reviews, fundraising, communitycreated encyclopedias, and more—they are every sort

of platform facilitating any sort of user expression,

including protected, lawful expression that is vulnerable

to a takedown demand.

Even when such accusations are sent by mistake the

censorial effect is still destructive to lawful expression,

which has now incur red an undeser ved punitive

consequence. Urban, at 96-97. It is even destructive

to the interests of creatives and copyright holders

themselves, because it means that there is nothing to

stop someone from taking down their own expressive

works. See, e.g., Timothy Geigner, Bungie, YouTuber

Settle Lawsuit Over Fraudulent DMCA Takedowns

For YouTube Videos, Techdirt (Mar. 29, 2024), https://

www.techdirt.com/2024/03/29/bungie-youtuber-settlelawsuit-over-fraudulent-dmca-takedowns-for-youtubevideos/. But the bigger threat to lawful expression comes

from when aspiring censors deliberately use takedown

notices to target speech they disfavor and cause it to be

disappeared. See, e.g., Stephanie Kirchgaessner, Top FDA

official demands removal of YouTube videos in which he

criticized Covid vaccines, The Guardian (Aug. 31, 2025),

https://w w w.theguardian.com/us-news/2025/aug/31/

11. Also affected are platforms using the 512(d) safe harbor,

which help that expression be found.

fda-official-youtube-videos. Or, worse, cause a speaker

to be silenced. See, e.g., Andy Maxwell, TV Museum

Will Die in 48 Hours Unless Sony Retracts YouTube

torrentfreak.com/tv-museum-will-die-in-48-hours-unlesssony-retracts-youtube-copyright-strikes-230904/. See also

Urban, at 47 (explaining “strikes,” as a mechanism some

platforms use to track infringement accusations against

users for termination purposes). In fact, if platforms

must terminate users because they’ve received multiple

accusations, there may be no need to send multiple

takedown notices to quiet a speaker because even one

can cow their target into removing their own potentially

lawful expression, no matter how non-infringing, in order

to avoid accruing a “strike” and putting their ability to

speak through the service at all at risk. Given that getting

terminated can mean getting cut off from long-cultivated

audiences and potentially even livelihoods, it is a risk

many speakers can ill afford to take. See discussion infra

Section I.B.

While takedown notices are supposed to be sent

“under penalty of perjury,” 17 U.S.C. § 512(c)(3)(A)(vi),

and in theory the DMCA provides a remedy when they

are wrongful, 17 U.S.C. § 512(f), in practice it is near

impossible for someone whose lawful speech has been

unduly removed by an invalid takedown notice to obtain

a remedy. See, e.g., Lenz v. Universal Music Corp., 815

F. 3d 1145, 1153-1156 (9th Cir. 2015), cert. denied, 137 S.

Ct. 2263 (2017) (allowing but nevertheless frustrating a

mom’s recovery under 512(f) for having had her lawful

speech—a video of her child dancing to a fair use of

music—removed via an invalid takedown notice). Because

senders know their accusations can have this censorial

effect, with little practical risk to themselves, there is little

to deter them from sending illegitimate notices and lawful

speech frequently and inevitably ends up undeservedly

censored. See Eric Goldman, In 512(f), the “F” Stands

for “Futility”, Technology & M arketing Law Blog (Aug.

15, 2025), https://blog.ericgoldman.org/archives/2025/08/

in-512f-the-f-stands-for-futility-shaffer-v-kavarnos.htm.

Nevertheless, despite the epidemic of invalid

takedown notices, the theory of liability raised in these

Cox cases requires platforms to automatically treat all

these accusations as inherently legitimate. Every concern

about “knowledge” frets that Cox knew users were

infringing and yet did nothing, with all that “knowledge”

predicated on the notices it received. The theory further

suggests that if the platform knew of a past infringement it

was obligated to presume to know that the next thing the

user would do would also be infringing. But acting upon

such presumptions is the essence of prior restraint, where

future expression is restricted as a result of a past wrong.

Near, 283 U.S. at 713. Even where a past wrong may have

actually been committed such a prior restraint on future

expression is still unconstitutional. Id. at 720 (“He does

not lose his right by exercising it.”). But the expressive

harm is even more severe where, as here, the past act was

only allegedly wrongful, and especially when, as here,

the conclusion that knowledge of infringement was truly

imparted via these notices cannot withstand scrutiny. As

BMG v. Cox evinced, all Cox actually had knowledge of

was that there was an accusation of infringement. But it

did not know, nor could it know,12 whether there actually

was any infringement—and, as all those dismissed claims

12. See also Brief for Petitioner at 10 (observing that it had

no way of knowing, even presuming an infringement accusation

were valid, which user committed it).

showed, in many cases there was not. Yet Cox was still

expected to act as though there were, and under risk of

great penalty if it did not.

Although on occasion platforms do refuse to act upon

certain takedown notices that they become aware of

and suspect to be invalid, given the sheer number they

receive13 they can defend only the smallest proportion

of their users and an even smaller proportion of their

expression,14 and they attempt it at great legal peril

because refusing a takedown notice can jeopardize their

safe harbor protection and risk shared liability. Having

an “opportunity” to defend against the accusations also

does not avoid the prior restraint problem, because a

consequence has already accrued if that defense needs to

be made to forestall further penalty. See Near, 283 U.S.

at 721.15

13. See id. at 8 (“By 2008, the music industry was flooding

ISPs with millions of notices a year.”).

14. See Moody, 144 S. Ct. at 2395 (“The biggest socialmedia companies—entities like Facebook and YouTube—host a

staggering amount of content. Facebook users, for example, share

more than 100 billion messages every day. And YouTube sees more

than 500 hours of video uploaded every minute.”).

15. The DMCA includes a “putback” provision allowing

users whose expression is removed to defend it to the platform.

17 U.S.C. § 512(g)(2). Yet not only is it inadequate doctrinally to

still demand an accused user defend himself within this system

of extrajudicial censorship, but history has also shown it to be

inadequate practically to defend against wrongful takedown

demands. Urban, at 44-46. Furthermore, it applies only to formal

takedown demands for the 512(c) safe harbor and not the informal

accusations made under the 512(a) safe harbor that never result

in anything being “taken down.”

Additionally, as a practical matter, platforms are also

generally ill-equipped to know, or expediently discover,

who the legitimate copyright holder is, whether the

notice has been sent by an authorized agent, or whether a

challenged use of a work is either licensed or fair. In fact,

the DMCA itself notably excuses platforms from having

to police their users’ activity themselves as a condition of

their safe harbor protection, as they are not in a position to

do so effectively. 17 U.S.C. § 512(m)(1). But that provision,

along with the safe harbor system itself, are all obviated

when the only way for a platform to protect itself is

to presume every accusation brought to it is valid and

automatically act against the user and their expressive

activity, regardless of whether the user’s online activity

was lawful and despite it may well being. See Bantam,

372 U.S. at 70 (finding a legal regime untenably hostile to

protected expression when it lacked safeguards to ensure

that non-wrongful, protected expression would not be

suppressed too).

The resulting speaker terminations, and 512(c)

speech removals, all punitive consequences aimed at

suppressing expression, see Near, 283 U.S. at 711, are all

restraints upon expressive rights prior to a proper judicial

determination that such consequence is warranted.

Likewise is the looming secondary liability causing Cox

to apply these sanctions, as both a prior restraint on Cox

itself, to force it to act in the wake of an accusation, see

Bantam, 372 U.S. at 67-68 (“Silverstein’s compliance with

the Commission’s directives was not voluntary.”), and

on the affected users. As this Court has recently found,

applying legal pressure to a third party as a means of

affecting another indirectly is legally no different than

applying that legal pressure to them directly. Nat. Rifle

Association of America v. Vullo, 144 S. Ct. 1316, 1322

(2024); see also discussion infra Section I.C. But when

it happens without proper adjudication first it is, like in

Bantam, “a scheme of state censorship effectuated by

extralegal sanctions” and thus an unconstitutional prior

restraint. Bantam, 372 U.S. at 71-72.

B. There is a Packingham problem

Even when infringement accusations are valid the

extrajudicial remedy Cox was expected to apply if it

wanted to avoid liability—terminating users—is not. Nor

could it have been even if it were a penalty imposed by law

directly on the user after sufficient due process.

As this Court found in Packingham v. North Carolina,

being kept off the Internet is too severe a penalty even for

adjudicated criminals who have caused tangible harm to

the physical well-being of their victims. 137 S. Ct. 1730,

1737 (2017) (“Even convicted criminals—and in some

instances especially convicted criminals—might receive

legitimate benefits from these means for access to the

world of ideas, in particular if they seek to reform and to

pursue lawful and rewarding lives.”). Meanwhile copyright

infringement is not even a criminal offense except

in limited circumstances inapplicable to this or most

other cases involving Internet platforms or their users.

See 17 U.S.C. § 506. And to the extent it causes harm,

that harm is only economic, and itself debatable. Mike

Masnick, The Sky Is Rising 2024 Edition: Rather Than

Destroying Culture, The Internet Has Saved The Content

Industries, Techdirt (Jan. 22, 2024) https://www.techdirt.

com/2024/01/22/the-sky-is-rising-2024-edition-ratherthan-destroying-culture-the-internet-has-saved-the-

content-industries/. To allow such a penalty here, one that

this Court already deemed intolerable in circumstances

far more consequential, would yield the absurd result of

effectively treating copyright infringement as a worse

offense than some of the world’s most heinous crimes.

Packingham, 137 S. Ct. at 1736 (“The sexual abuse of a

child is a most serious crime and an act repugnant to the

moral instincts of a decent people.”).

But even if the same reasoning of Packingham were

to be applied afresh to the copyright context, and not

just comparatively, forcing platforms to terminate users

would still fail to survive even intermediate scrutiny. Id.

To survive intermediate scrutiny, a law must be “narrowly

tailored to serve a significant governmental interest.” Id.

The goal of copyright law is to promote the progress of

science and art. U.S. Const. art. I, § 8, cl. 8. But even to

the extent that such promotion is a significant interest for

the government to pursue, driving away the speakers and

speech who were contributing to it is entirely inconsistent

with that policy value. The regulatory means of account

terminations are simply not narrowly tailored enough to

survive scrutiny because termination does not just target

infringing behavior but also non-infringing. Packingham,

137 S. Ct. at 1736 (“[T]he law must not ‘burden substantially

more speech than is necessary to further the government’s

legitimate interests.’”). Once users are terminated they

cannot engage in any sort of expressive activity, no matter

how lawful. Id. at 1737 (“[T]o foreclose access to social

media altogether is to prevent the user from engaging in

the legitimate exercise of First Amendment rights.”). And

this Court has already found such a hit to speech rights,

even to those of convicted criminals, to be a hit too hard.

Id. (“Even convicted criminals—and in some instances

especially convicted criminals—might receive legitimate

benefits from these means for access to the world of ideas,

in particular if they seek to reform and to pursue lawful

and rewarding lives.”).

In Packingham this Court was also contending

with a law that ostensibly targeted the use of a subset

of Internet services, namely “social media.” Id. at 17361737. But it recognized that the law potentially had the

effect of putting all of the riches of the Internet beyond

the reach of certain people. Id. at 1737 (“By prohibiting

sex offenders from using those websites, North Carolina

with one broad stroke bars access to what for many

are the principal sources for knowing current events,

checking ads for employment, speaking and listening in

the modern public square, and otherwise exploring the

vast realms of human thought and knowledge.”). The case

against Cox illustrates how the termination provision

threatens the same breadth of offerings because Cox

is a general purpose ISP. While it would still be plenty

problematic to cut users off from the various specialized

services, including those eligible for the 512(c) safe

harbor like “social media,”16 cutting them off from fullservice broadband ISPs causes even more serious harm

to speakers because the broadband ISP market is not a

particularly competitive one. See Mike Masnick, Just A

Click Away: How To Improve Broadband Access, Techdirt

(Jul. 18, 2022), https://www.techdirt.com/2022/07/18/just16. It is a live concern that this case will bear on. See, e.g.,

Concord Music Group, Inc. v. X Corp., No. 3:23-cv-00606, slip op.

at 18 (M.D. Tenn. Mar. 5, 2024) (order denying defendant’s motion

to dismiss a contributory liability claim partly on the basis of the

provider allegedly not having terminated repeat infringers).

a-click-away-how-to-improve-broadband-competition/.

Users may effectively have only one choice of provider, if

they want to be able to use the Internet at all. See, e.g., Karl

Bode, Two Decades Later And The FCC Is Still Trying

To Crack Down On Anti-Competitive Deals Between

Landlords And Broadband Monopolies, T echdirt

(Mar. 18, 2024), https://www.techdirt.com/2024/03/18/

two-decades-later-and-the-fcc-is-still-trying-to-crackdown-on-anti-competitive-deals-between-landlords-andbroadband-monopolies/. Consequently, if they lose access

to that provider’s services then they may lose access to the

Internet altogether. Given how ubiquitously important the

Internet is, see, e.g., Packingham, 137 S. Ct. at 1735-1736,

losing online access is a particularly injurious consequence

to a speaker and disproportionate to the harm alleged,

even one that might be deserving of some sort of penalty.

And it is completely untenable to inflict such a penalty

on a speaker who is not deserving, not just on the basis of

unfounded accusations but possibly no accusation at all.

As Cox also points out, cutting off a “user” may involve

cutting off what is essentially a service provider itself,

terminating entire households full of users, hospitals,

hotels, coffee shops, universities, military barracks,

or even regional ISPs providing service to potentially

thousands or tens of thousands of their own users. Brief

for Petitioner at 9. Were these accounts to be terminated

it would mean not just one speaker being denied the

ability to speak anymore via the platform but multiple,

and potentially countless, and based on nothing that

they themselves have done. There is unlikely to be any

legitimate government interest that such an injustice could

possibly advance.

There is a Vullo problem

Every speaker on the Internet is dependent on other

systems and services to get their expression online. See

discussion infra Section II.C. As a result, when a state

actor wants to target their speech it may seem more

expedient to pressure the services Internet speakers

depend on, to cause the services to act against those

speakers or their speech, rather than it trying to reach the

speakers or speech directly, which may be less practical,

if not constitutionally forbidden outright.

But this tactic is just as unconstitutional as what this

Court deemed so in Vullo. There this Court found that the

government violated the NRA’s First Amendment rights

by exploiting its dependency on other parties and using

pressure on those other parties as a means of pressuring

it. Id. at 1331 (“The analogy is to killing a person by

cutting off his oxygen supply rather than shooting him.”).

Targeting a speaker did not suddenly become constitutional

just because it was applied to a middleman as opposed to

the target the state actor was trying to reach. Id. at 1326.

Yet the same constitutional infirmity with this regulatory

approach applies here. Just as a state actor there could

not pressure an intermediary another depended on as a

means of affecting that other’s expression, neither can

it impose laws that pressure Internet intermediaries to

interfere with online expression. The First Amendment

forbids it, for any purpose—even copyright.17

17. It makes no difference if the user were actually culpable

because, without a judicial finding, no one actually knows whether

the user is indeed culpable. And this regulatory model, of

pressuring the platform to sanction the user instead of a court,

means that no one will ever know before the sanction is applied.

This regulatory model also affects the rights of the

platform too, which has its own First Amendment rights.

See Near, 283 U.S. at 720 (finding that these rights could

not be ignored because the regulatory target was in the

“business” of facilitating the publication of the allegedly

objectionable material). Ordinarily the editorial discretion

platforms exercise in choosing what speakers and speech

to associate with is protected by the First Amendment,

just as traditional press outlets’ choices have been.

Moody, 144 S. Ct. at 2405 (2024). But that discretion is

lost when platforms are forced by legal pressure to make

certain decisions with respect to their users or their

users’ expression. While the government can attempt to

persuade platforms to act in certain ways, when it uses

its power coercively, the constitutional problem arises.

See Murthy v. Missouri, 144 S. Ct. 1972, 2010 (2024)

(Alito, J., dissenting) (citing Vullo, 144 S. Ct. at 1326).

Such coercion is present here. Per the DMCA, at least as

the lower courts here interpreted it, Cox had to “decide”

to terminate its users if it wanted to retain its statutory

defense against liability. So would all other platforms,

including those using the 512(c) safe harbor, who would

also need to remove any other user expression accused

of being infringing. 17 U.S.C. § 512(c)(1)(C). But when

the decision to remove speakers or speech is an offer

they effectively cannot refuse without risking ruinous

consequences then it is not much of a choice after all.

The conditional nature of the DMCA, which forces

platforms make the “choice” to act against speakers or

speech, or else risk potentially enormous liability, creates

an equally enormous legal pressure subordinating their

First Amendment-protected editorial and associative

discretion. But the real culprit taking away that discretion

is the threat of secondary liability itself and the risk of

having to be responsible for what their users do. This

looming liability is the real gun to platform providers’

heads, because losing the safe harbor would not matter

so much if the consequences of losing it were not so dire.

It is this legal pressure that ultimately forces them to act

against their users and their expression and thus creates

a constitutional problem.

II. A legal regime that could hold Cox liable would

affect all online expression

This case involves liability for copyright infringement.

And it arises from conduct—filesharing—that may often

be infringing. But that the accused users themselves may

in fact be culpable should not distract from what is at issue

in this case. This case is not about holding Internet users

themselves liable for their own actions; it is about holding

someone else liable for them. But if such secondary liability

can be tolerated here, for copyright, it can be tolerated

for any other form of alleged liability targeting any sort

of platform, and not just 512(a) conduit providers like Cox,

but all Internet platforms. The practical consequence of

such a liability model will be the chilling of all Internet

expression, no matter how lawful or beneficial, by chilling

the platforms that enable it.

Targeting platforms with secondary liability

inherently affects user expression

Although the way the DMCA is currently interpreted

has some constitutional infirmities, the basic idea of

providing platforms protection from liability arising

from how their users used their services is nevertheless

an important one, for the same reason that “Section 230”

is also important. See 47 U.S.C. § 230(c); see also Brief

amicus curiae of M. Chris Riley, Floor64, Inc. d/b/a

The Copia Institute, and Engine Advocacy (“Gonzalez

Amicus”), Gonzalez v. Google, 598 U.S. 617 (2023) (211333). The Internet is a communications medium that

inherently depends on intermediaries to help expression

get from one person to another. These intermediaries

come in all sorts of shapes and sizes. The four DMCA safe

harbors catalog several buckets that they may fall into,

but within each category the types of service providers

are myriad, ranging from providers of high-capacity

backbone infrastructure, to ISPs like Cox that serve as

the onramps for users to connect to the wider Internet,

to other forms of online services that help users express

themselves and interact with others’ online expression,

which can include email, social media, marketplaces,

streaming services, or online storage, to name just a few.

Together they allow people to engage with expression in

infinite ways, be it contemporaneously or asynchronously,

via one-to-one communications or one-to-many, and via

stored information or live exchange. And a fundamental

reality is that no Internet communication could happen

at all but for these providers being available, and able, to

help facilitate that exchange of expression.

But it is also a fundamental reality that few if any

intermediaries could be available and able to help

facilitate the exchange of expression were it not legally

safe to do that helping. Whether it is a statute or doctrinal

limitation on secondary liability providing that safety,

that protection is still crucial. If platforms could be

forced to share in whatever liability might be manifest

in the expression they help facilitate—or even simply

have to answer for accusations of liability—it would be

too risky to help facilitate very much of it, or even any

of it. Zeran v. America Online, Inc., 129 F.3d 327, 331

(4th Cir. 1997). Even the smallest providers can facilitate

expression in enormous amounts, especially in proportion

to their own size and resources. Intermediary platforms

are not necessarily all large, well-capitalized companies;

intermediary services can even be offered by non-profits

or even individuals, like those with their own blogs that

allow for user interaction in comments, or those who run

their own social media instances. See Gonzalez Amicus

at 1-2.

Furthermore, it is not just ultimate liability that is

an issue for intermediaries; even if there were no liability

manifest in the user content, simply having to defend

against accusations of liability, even when non-meritorious,

can itself be ruinous. See, e.g., UMG Recordings, Inc.

v. Shelter Capital Partners, 718 F. 3d 1006, 1011 (9th

Cir. 2013) (holding the platform non-liable after it was

bankrupted by the defense); Mike Masnick, Veoh Still

Not Dead Enough For Universal Music; Asks Court

To Rehear Case Yet Again, T echdirt (Feb. 11, 2010),

https://www.techdirt.com/2013/04/08/veoh-still-not-deadenough-universal-music-asks-court-to-rehear-case-yetagain/. The cost can be tremendous to defend a single

accusation, let alone multitudinous ones arising from any

sort of volume of user activity.18 Statutes like the DMCA,

18. Engine, Startups, Content Moderation, & Section

2 30 at 4 - 5, available at https://static1.squarespace.com /

static/571681753c44d835a440c8b5/t/61b26e51cdb21375a31d31

2f/1639083602320/Startups%2C+Content+Moderation%2C+an

d+Section+230+2021.pdf.

or the less-conditional Section 230,19 are intended to

provide platforms with the safety they need to be in the

business of facilitating expression. Shelter Capital, 718

F. 3d at 1014; Zeran, 129 F.3d at 330-331. Without that

safety platforms’ ability to perform that facilitation will be

compromised. See Fair Housing Coun. Of San Fernando

Valley v. Roommates. Com, LLC, 521 F. 3d 1157, 1174 (9th

Cir. 2008) (observing that, without protection, platforms

would face “death by ten thousand ducks bites, fighting

off claims that they promoted or encouraged—or at least

tacitly assented to—the illegality of their users.”). They

would need to police the nearly infinite and infinitely varied

volume of user expression, which is an impossible task.

Already platforms cannot even reliably know whether user

expression is infringing; they are certainly not going to be

able to know if it potentially violates all other law, arising

from potentially infinite jurisdictions. Putting a liability

onus on them means that they will have no choice but to

suppress much of it, if not turn it away outright.

This resulting suppression of online activity is often

also not just a collateral effect; in many instances it is the

goal of the regulatory scheme. Allowing it here would

greenlight the practice of regulating online speech via

middleman pressure. See discussion supra Section I.C.

19. Section 230 protects platforms from most forms of

liability. 47 U.S.C. § 230(c)(1) (“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.”). Although its protection is not conditional on platforms

satisfying any particular criteria, as the DMCA’s is, it does not

apply to “intellectual property” liability, 47 U.S.C. § 230(e)(2),

which is instead in the purview of the DMCA when it comes to

Such an outcome would be bad even when the potential

liability in the user activity is something that can

constitutionally be deemed wrongful, because, even as in

this filesharing case, there is still a due process problem

with people suffering a punitive consequence for their

online activity without proper adjudication. But it is even

worse when it is used as a way of regulating user speech

in ways where the Constitution would prevent it. See, e.g,,

Mike Masnick, Trump Promises To Abuse Take It Down

Act For Censorship, Just As We Warned, Techdirt (Mar.

6, 2025), https://www.techdirt.com/2025/03/06/trumppromises-to-abuse-take-it-down-act-for-censorshipjust-as-we-warned/ (describing how a law ostensibly

intended to cause platforms to moderate non-consensual

imagery was ripe to be weaponized as a means to force

them to take down political speech). See also Mike

Masnick, Amy Klobuchar Wants To Break The Internet

Because Someone Made A Stupid Satirical Video About

Her, T echdirt (Aug. 21, 2025), https://www.techdirt.

com/2025/08/21/amy-klobuchar-wants-to-break-theinternet-because-someone-made-a-stupid-satirical-videoabout-her/ (describing a legislative proposal that would

make it impossible for platforms to allow AI imagery to be

used in protected expression); Mike Masnick, Bipartisan

Group Of Senators Introduce New Terrible ‘Protect The

Kids Online’ Bill, Techdirt (May 2, 2024), https://www.

techdirt.com/2024/05/02/bipartisan-group-of-senatorsintroduce-new-terrible-protect-the-kids-online-bill/

(describing a legislative proposal that would affect young

people’s ability to exercise their First Amendment rights

online by forcing platforms to keep them off); Cathy

Gellis, Senator Klobuchar Proposes An Unconstitutional

Law That Would Kill Legions Of People If Trump Were

Still President, Techdirt (Jul. 26, 2021), https://www.

techdirt.com/2021/07/26/senator-klobuchar-proposesunconstitutional-law-that-would-kill-legions-people-iftrump-were-still-president/ (describing a legislative

proposal to force platforms to moderate health information

as policymakers preferred); Cathy Gellis, Creating State

Action Via Antitrust Law And Making The People

Who’ve Been Wrong About The Constitutionality Of

Content Moderation Suddenly Right, T echdirt (Jun.

28, 2021), https://www.techdirt.com/2021/06/28/creatingstate-action-via-antitrust-law-making-people-whovebeen-wrong-about-constitutionality-content-moderationsuddenly-right/ (describing past legislative proposals to

break up platform companies that do not moderate as the

government prefers).

As a regulatory practice, secondary liability for

platforms is also self-defeating. Platforms are already

working to solve the problems regulation cares about.

Many have entire Trust & Safety departments dedicated

to the task and there are conferences, trade groups,

and software companies to support them. See, e.g., Mike

Masnick, Trust & Safety Professional Association

Launches: This Is Important, Techdirt (Jun. 17, 2020),

at https://www.techdirt.com/2020/06/17/trust-safetyprofessional-association-launches-this-is-important/.

There is an entire industry built up around helping

platforms institutionally innovate solutions to some of the

toughest human problems as they are reflected online.

Toby Shulruff, Jeff Lazarus, and Amanda Menking, Voices

of Trust & Safety: Origins and Evolutions (2025), https://

papers.ssrn.com/sol3/papers.cfm?abstract_id=5361961.

Meanwhile individual human operators of small platforms

also do the best they can, exercising their own personal

rights to editorial and associative discretion. See Gonzalez

Amicus at 24. As Congress understood when it passed

Section 230, to get the best regulatory results—the most

beneficial expression online, and the least deleterious—

platforms need to be free to moderate their platforms,

associating with and disassociating with speakers and

speech, as they see fit. Zeran, 129 F. 3d at 331. The

DMCA too, even though it encroaches upon that protected

discretion, was still intended to provide platforms with

at least some protection so that they could remain in the

business of intermediating expression while also removing

infringing material without risking liability for these

efforts. Shelter Capital, 718 F. 3d at 1014. Ultimately the

most effective regulatory schemes are those that make it

possible for platforms to solve the problems we want them

to solve, either themselves or via a competitive landscape

where more and better platforms can always come into

existence. Secondary liability against them inevitably

prevents that potential and always comes at the cost of

the expressive function platforms are needed to deliver.

B. There is nothing about copyright, filesharing,

or Cox that justifies secondary liability

Secondary liability for platforms is chilling to

the expressive rights of both them and their users. If

nevertheless it is allowed here, for Cox, for copyright,

for a case about filesharing, there is no limiting principle

preventing a similar liability regime from forcing any other

platform to answer for any other accused wrongdoing

arising from their users. The same constitutional defects

present here could arise in potentially infinite contexts.

Part of that danger is due to there being nothing so

exceptional about copyright that would justify forgiving

the constitutional defects in a case like this one, and only

in a case like this one. Filesharing itself, even if often

done for infringing purposes, would still not warrant it.

At essence “filesharing” is simply the technical means

to efficiently exchange information online, and like so

many other technologies helping informational content be

created and consumed, it is also used for non-infringing

purposes. See Metro-Goldwyn-Mayer Studios Inc. v.

Grokster, Ltd., 545 U.S. 913, 942 (2005) (citing Sony Corp.

of America v. Universal City Studios, Inc., 464 U.S. 417,

442 (1984)). Copyright law itself must also co-exist with

the First Amendment, Eldred v. Ashcroft, 537 U.S. 186,

219 (2003), which prohibits prior restraint. Concerns

about infringement, however apt, cannot justify playing

with a different set of constitutional rules, especially not

when straying from them inevitably leads to censorship—

including of the very creative works copyright law is

supposed to incentivize. The constitutional authority for

having copyright law at all is granted for the purpose of

promoting the exchange of ideas and information; when

the resulting law instead has a smothering effect on that

exchange, it is no longer meeting its constitutionallyprescribed purpose and cannot be tolerated, let alone

privileged.

Furthermore, it is copyright exceptionalism itself,

opening the door to expansive secondary liability, that

has allowed this self-defeating harm. Because Section 230

has effectively barred secondary liability for most other

liability contexts other than copyright, there has been a

steady stream of examples of attempts to shoehorn other

liability complaints into the copyright claims in order

to tap into the censorial power of this sort of copyright

liability regime. See, e.g., Peter Guest, Exposed documents

reveal how the powerful clean up their digital past using

a reputation laundering firm, Rest of World (Feb. 3,

2022), https://restofworld.org/2022/documents-reputationlaundering-firm-eliminalia/.

Continuing this sort of copyright exceptionalism

therefore raises at least two concerns. One is that

for holding platforms liable for their users’ expression. But

the other is that if an exception is carved out for copyright,

to justify its expansive secondary liability as distinct

from ordinary circumstance, that distinction may soon

disappear as a similar rationale gets employed in future

non-copyright matters as a path for finding secondary

liability despite the restraint ordinarily required by

common law doctrine. See, e.g., Twitter, Inc. v. Taamneh,

143 S. Ct. 1206, 1222 (2023). Instead of being the exception,

it may now become the rule.

This evolutionary dynamic is not a hypothetical

concern, as the sort of expansive secondary liability

regime at issue here has already become a tempting tool

for regulators to deploy in other contexts, where traditional

doctrine would have otherwise counseled restraint.

Section 230 currently does a lot of heavy lifting to keep

expansive secondary liability for non-copyright reasons at

bay, but Section 230 is a statute that Congress could repeal

or otherwise attempt to limit via other legislation, should

it choose to. What truly protects online speech are the

constitutional principles that prevent holding platforms

secondarily liable for their users’ expressive activity at all.

Unfortunately, applying pressure through secondary

liability has lately become an appealing regulatory model

to lawmakers everywhere, possibly because they see

it apparently working here, in the copyright context. A

great deal of brewing regulation, and some laws already

recently passed, 20 is built around applying a penalty to

platforms if users are able to use their services in ways

the lawmakers disfavor. But it is all still impermissible

pressuring of platforms as a means of affecting online

expression. All of these legislative proposals aimed

at platforms are built around the idea that the way to

control what people can say or do online is to squeeze the

intermediary they need to do it so that the intermediary

acts against them, without due process, and without the

government needing to try to target them directly. Each

one of these proposals is constitutionally suspect though,

for the same reasons that secondary liability here is. See

discussion supra Section I.A-C.

That the DMCA has nevertheless been chugging

along for nearly 30 years does not obviate the inherent

constitutional problem. For one thing, the problem has

gotten worse, and this Cox case is an example of that

worsening trend because of how the preceding lower

court decisions have reduced the protective utility of the

DMCA safe harbors. Furthermore, if the liability against

Cox is allowed to stand, it will fundamentally change the

legal universe in which platforms operate, making the

facilitation of anyone’s online activity a much more fraught

endeavor for all platforms to continue. As a result, many

will stop, thus depleting the ability for anyone to speak

online, or otherwise heavily constrain how they can be

used, which will lead to the same deleterious effect on

online expression. It will also not just be the broadband

providers affected, but any sort of platform, helping users

and expression to interact online in myriad ways.

20. See discussion supra Section II.a.

The principles espoused in Vullo have also not been

directly applied to platform liability before now. This case

is the first, but surely not the last, to put these sorts of

constitutional questions squarely before this Court. For the

previously described reasons, this Court should find that

this regulatory model, of holding platforms responsible

for their users’ activity, is inherently suspect, not just

here, in copyright cases, but in all cases. Exonerating it

here is only going to encourage it to spread, as is evident

from how this regulatory model has already proliferated

in so many legislative proposals in so many other liability

contexts. That it has been tolerated for so long appears

to have created the illusion that there is no significant

constitutional defect with it, when there always has been.

It is simply finally time to notice.

CONCLUSION

For the foregoing reasons, Cox should not be found

liable for the allegedly infringing acts of their users.

Respectfully submitted,

Catherine R. Gellis, Esq.

Counsel of Record

3020 Bridgeway, #247

Sausalito, CA 94965

(202) 642-2849

cathy@cgcounsel.com

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.