116 F.4th 180

3rd Cir.

2024

TAWAINNA ANDERSON, INDIVIDUALLY AND AS ADMINISTRATIX OF THE ESTATE OF N.A., A DECEASED MINOR v. TIKTOK, INC.; BYTEDANCE, INC.

No. 22-3061

UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT

August 27, 2024

PRECEDENTIAL. On Appeal from the United States District Court for the Eastern District of Pennsylvania (D.C. Civil No. 2-22-cv-01849). U.S. District Judge: Honorable Paul S. Diamond. Argued January 17, 2024.

Before: SHWARTZ, MATEY, and PHIPPS, Circuit Judges.

Jeffrey P. Goodman [ARGUED]

Robert J. Mongeluzzi

Saltz Mongeluzzi & Bendesky

1650 Market Street

One Liberty Place, 52nd Floor

Philadelphia, PA 19103

Counsel for Appellant Tawainna Anderson

Geoffrey M. Drake

King & Spalding

1180 Peachtree Street NE

Suite 1600

Atlanta, GA 30309

Albert Giang

King & Spalding

633 W 5th Street

Suite 1600

Los Angeles, CA 90071

David Mattern

King & Spalding

1700 Pennsylvania Avenue NW

Suite 900

Washington, DC 20006

Joseph O‘Neil

Katherine A. Wang

Campbell Conroy & O‘Neil

1205 Westlakes Drive

Suite 330

Berwyn, PA 19312

Andrew J. Pincus [ARGUED]

Nicole A. Saharsky

Mayer Brown

1999 K Street NW

Washington, DC 20006

Mark J. Winebrenner

Faegre Drinker Biddle & Reath

90 S. Seventh Street

2200 Wells Fargo Center

Minneapolis, MN 55402

Counsel for Appellees TikTok, Inc. and ByteDance, Inc.

OPINION OF THE COURT

SHWARTZ, Circuit Judge.

TikTok, Inc., via its algorithm, recommended and promoted videos posted by third parties to ten-year-old Nylah Anderson on her uniquely curated “For You Page.” One video depicted the “Blackout Challenge,” which encourages viewers to record themselves engaging in acts of self-asphyxiation. After watching the video, Nylah attempted the conduct depicted in the challenge and unintentionally hanged herself. Nylah‘s mother, Tawainna Anderson, sued TikTok and its corporate relative ByteDance, Inc., (collectively, “TikTok“) for violations of state law. The District Court dismissed her complaint, holding that the Communications Decency Act (“CDA“), 47 U.S.C. § 230, immunizes TikTok. For the following reasons, we will reverse in part, vacate in part, and remand.

I

A¹

TikTok is a video-sharing social media platform that allows users to create, post, and view content. TikTok users can search the platform for content or, without searching, view content that TikTok‘s algorithm recommends by posting the content to a user‘s “For You Page” (“FYP“).² TikTok‘s algorithm is not based solely on a user‘s online inputs. Rather, the algorithm curates and recommends a tailored compilation of videos for a user‘s FYP based on a variety of factors, including the user‘s age and other demographics, online interactions, and other metadata.

Some videos that may appear on users’ FYPs are known as “challenges,” which urge users to post videos of themselves replicating the conduct depicted in the videos. The “Blackout Challenge ... encourages users to choke themselves with belts, purse strings, or anything similar until passing out.” App. 31 (Compl. ¶ 64). TikTok‘s FYP algorithm recommended a Blackout Challenge video to Nylah, and after watching it, Nylah attempted to replicate what she saw and died of asphyxiation.

B

Anderson, as the administratrix of Nylah‘s estate, sued TikTok in the United States District Court for the Eastern District of Pennsylvania, asserting claims for, among other things, strict products liability and negligence.³ She alleges that TikTok: (1) was aware of the Blackout Challenge; (2) allowed users to post videos of themselves participating in the

Blackout Challenge; and (3) recommended and promoted Blackout Challenge videos to minors’ FYPs through its algorithm, including at least one such video to Nylah‘s FYP, which resulted in her death. The District Court dismissed the complaint, holding that TikTok was immune under § 230 of the CDA, 47 U.S.C. § 230. Anderson v. TikTok, Inc., 637 F. Supp. 3d 276, 282 (E.D. Pa. 2022).

Anderson appeals.⁴

II⁵

Congress enacted § 230 of the CDA to immunize interactive computer services (“ICSs“)⁶ from liability based on content posted by third parties in certain circumstances. See

F.T.C. v. Accusearch Inc., 570 F.3d 1187, 1195 (10th Cir. 2009).⁷ Section 230 immunizes ICSs only to the extent that they are sued for “information provided by another information content provider.”⁸ 47 U.S.C. § 230(c)(1).⁹ In other words, ICSs are immunized only if they are sued for someone else‘s expressive activity or content (i.e., third-party speech), but they are not immunized if they are sued for their own expressive activity or content (i.e., first-party speech).

Anderson asserts that TikTok‘s algorithm “amalgamat[es] [] third-party videos,” which results in “an expressive product” that “communicates to users ... that the curated stream of videos will be interesting to them[.]” ECF No. 50 at 5. The Supreme Court‘s recent discussion about algorithms, albeit in the First Amendment context, supports

this view.¹⁰ In Moody v. NetChoice, LLC, 144 S. Ct. 2383 (2024), the Court considered whether state laws that “restrict the ability of social-

media platforms to control whether and how third-party posts are presented to other users” run afoul of the First Amendment. 144 S. Ct. 2383, 2393 (2024). The Court held that a platform‘s algorithm that reflects “editorial judgments” about “compiling the third-party speech it wants in the way it wants” is the platform‘s own “expressive product” and is therefore protected by the First Amendment. Id. at 2394.

Given the Supreme Court‘s observations that platforms engage in protected first-party speech under the First Amendment when they curate compilations of others’ content via their expressive algorithms, id. at 2409, it follows that doing so amounts to first-party speech under § 230, too. See Doe ex rel. Roe v. Snap, Inc., 144 S. Ct. 2493, 2494 (2024) (Thomas, J., dissenting from denial of certiorari) (observing that “[i]n the platforms’ world, they are fully responsible for their websites when it results in constitutional protections, but the moment that responsibility could lead to liability, they can disclaim any obligations and enjoy greater protections from suit than nearly any other industry.“).

Because TikTok concedes that Anderson‘s complaint “describe[s] an algorithm indistinguishable from those addressed in NetChoice[,]” ECF No. 51 at 2, which the Supreme Court described as one that results in expressive speech, NetChoice, 144 S. Ct. at 2405 (holding that “social-media platforms are in the business, when curating their feeds, of combining multifarious voices to create a distinctive expressive offering” (internal quotation marks and citation omitted)), we need not weigh in on whether other algorithms result in expressive speech. Moreover, because TikTok‘s “algorithm, as described in the complaint, does not” “‘respond solely to how users act online,‘” ECF No. 51 at 2 (quoting NetChoice, 144 S. Ct. at 2404 n.5), TikTok makes choices about the content recommended and promoted to specific users, and by doing so, is engaged in its own first-party speech.

Here, as alleged, TikTok‘s FYP algorithm “[d]ecid[es] on the third-party speech that will be included in or excluded from a compilation—and then organiz[es] and present[s] the included items” on users’ FYPs. NetChoice, 144 S. Ct. at 2402. Accordingly, TikTok‘s algorithm, which recommended the Blackout Challenge to Nylah on her FYP, was TikTok‘s own “expressive activity,” id., and thus its first-party speech. Such first-party speech is the basis for Anderson‘s claims. See App. 39 (Compl. ¶¶ 107(k), (o)) (alleging, among other things, that TikTok‘s FYP algorithm was defectively designed because it “recommended” and “promoted” the Blackout Challenge).¹¹ Section 230 immunizes only information “provided by another[,]” 47 U.S.C. § 230(c)(1), and here, because the information that forms the basis of Anderson‘s lawsuit—i.e., TikTok‘s recommendations via its FYP algorithm—is TikTok‘s own expressive activity, § 230 does not bar Anderson‘s claims.¹²^,¹³

III

For the foregoing reasons, we will reverse in part, vacate in part, and remand.¹⁴

MATEY, Circuit Judge, concurring in the judgment in part and dissenting in part.

TikTok reads § 230 of the Communications Decency Act, 47 U.S.C. § 230, to permit casual indifference to the death of a ten-year-old girl. It is a position that has become popular among a host of purveyors of pornography, self-mutilation, and exploitation, one that smuggles constitutional conceptions¹ of a “free trade in ideas” into a digital “cauldron of illicit loves” that leap and boil with no oversight, no accountability, no remedy.² And a view that has found support in a surprising number of judicial opinions dating from the early days of dial-up to the modern era of algorithms, advertising, and apps.

But it is not found in the words Congress wrote in § 230, in the context Congress acted, in the history of common carriage regulations, or in the centuries of tradition informing the limited immunity from liability enjoyed by publishers and distributors of “content.” As best understood, the ordinary meaning of § 230 provides TikTok immunity from suit for hosting videos created and uploaded by third parties. But it does not shield more, and Anderson‘s estate may seek relief for TikTok‘s knowing distribution and targeted recommendation of videos it knew could be harmful. Accordingly, I concur in the judgment in part and dissent in part.

I.

A.

Ten-year-old Nylah Anderson died after attempting to recreate the “Blackout Challenge” she watched on TikTok. The Blackout Challenge—performed in videos widely circulated on TikTok—involved individuals “chok[ing] themselves with belts, purse strings, or anything similar until passing out.” App. 31.³ The videos “encourage[d]” viewers to record themselves doing the same and post their videos for other TikTok users to watch. App. 31. Nylah, still in the first year of her adolescence, likely had no idea what she was doing or that following along with the images on her screen would kill her. But TikTok knew that Nylah would watch because the company‘s customized algorithm placed the videos on her “For You Page”⁴ after it “determined that the Blackout Challenge was ‘tailored’ and ‘likely to be of interest’ to Nylah.” App. 31.

No one claims the videos Nylah viewed were created by TikTok; all agree they were produced and posted by other TikTok subscribers. But by the time Nylah viewed these videos, TikTok knew that: 1) “the deadly Blackout Challenge was spreading through its app,” 2) “its algorithm was specifically feeding the Blackout Challenge to children,” and 3) several children had died while attempting the Blackout Challenge after viewing videos of the Challenge on their For You Pages. App. 31–32. Yet TikTok “took no and/or completely inadequate action to extinguish and prevent the spread of the Blackout Challenge and specifically to prevent the Blackout Challenge from being shown to children on their [For You Pages].” App. 32–33. Instead, TikTok continued to recommend these videos to children like Nylah.

B.

Following her daughter‘s death, Tawainna Anderson sued TikTok and its parent company, ByteDance, Inc. Anderson seeks to hold TikTok liable for 1) hosting the Blackout Challenge videos on its platform, 2) continuing to distribute the videos after it learned about the videos and the deaths that followed, and 3) recommending the videos to Nylah after TikTok knew the videos were likely to cause harm. TikTok moved to dismiss, arguing that Anderson sought to hold TikTok liable for acts completely immunized by § 230(c)(1). The District Court agreed.

II.

TikTok maintains that Anderson‘s claims are foreclosed by a nearly-limitless interpretation of § 230 adopted by several courts. But the best reading of the statute suggests a far narrower understanding of § 230 immunity.

A.

Like any man-made law, § 230 did not appear in a vacuum, and “some context is key to understanding Congress‘s aim” and the precise language it selected. OI Eur. Grp. B.V. v. Bolivarian Republic of Venez., 73 F.4th 157, 166 (3d Cir. 2023), cert. denied, 144 S. Ct. 549 (2024); see also 1 William Blackstone, Commentaries *61, *87. Congress enacted § 230 following more than a century of state and federal law regulating the transmission of third-party information and against the backdrop of two widely discussed judicial decisions addressing the liability of online service providers. Those decisions tracked traditional liability regimes that shielded parties who merely sent along allegedly harmful information, while imposing duties on those who did so with specific knowledge of the harmful nature of the content.

1. Begin with the birth of long-distance communication. Like the chat rooms and bulletin boards provided by 1990s online service providers, telegraph companies long served as the conduit for communication for much of the late nineteenth and early twentieth centuries. Given the immense market power of the telegraph,⁵ the law regularly imposed access and nondiscrimination duties familiar to physical networks like railroads.⁶ That raised questions about liability, since state laws often held companies responsible for negligent deliveries. See Adam Candeub, The Common Carrier Privacy Model, 51 U.C. Davis L. Rev. 805, 810–15 (2018). Liability could also attach based on the content of third-party information. See Adam Candeub, Reading Section 230 as Written, 1 J. of Free Speech L. 139, 145–47 & 146 n.26 (2021). While telegraph operators were ordinarily not responsible for the materials they transmitted, see O‘Brien v. W. U. Tel. Co., 113 F.2d 539, 541–43 (1st Cir. 1940), liability could attach if the company knew the content was harmful, see Von Meysenbug v. W. U. Tel. Co., 54 F. Supp. 100, 101 (S.D. Fla. 1944); see also Biden v. Knight First Amend. Inst. at Columbia Univ., 141 S. Ct. 1220, 1223 & n.3 (2021) (Thomas, J., concurring).⁷ But that was the rare exception.

This was the common-sense system throughout the twentieth century.⁸ Transmitters usually had little control over what rode their networks⁹ and rarely knew the circumstances that might make a statement harmful.¹⁰ Imposing liability for conduct that lacked culpability would unfairly punish beneficial industries and pin emerging networks under the weight of endless lawsuits. But the scale tipped in a different way when a transmitter of third-party information knew the content was harmful, a distinction that carried into the changes in communications technology during the back half of the twentieth century.

2. The internet began infiltrating daily life in the early 1990s through large commercial service providers like CompuServe, Prodigy, and AOL.¹¹ These emerging services “were born serving content of their own,”¹² but, facing competition, they expanded to allow “users to post comments on bulletin boards, open to other members, and to communicate in chat rooms.”¹³ Those added functions resurrected the old legal question familiar to common carriers: Should online service providers be liable for the actions of third parties on their networks? Understanding how courts answered this question is essential to understanding the legal context in which § 230 was enacted. Because a 1991 district court decision set the boundaries of liability law for the next three decades.

Believed to be the first case in the United States “to decide whether an online service ... could be held liable for third-party content,”¹⁴ Cubby, Inc. v. CompuServe, Inc. involved a defamation claim arising out of an allegedly libelous statement appearing on one of CompuServe‘s “special interest ‘forums.‘” 776 F. Supp. 135, 137 (S.D.N.Y. 1991). These fora, “comprised of electronic bulletin boards, interactive online conferences, and topical databases,” allowed

subscribers to post their own messages and interact with other users. Id. Pivoting from the closed curation of the old networks, CompuServe did not review subscriber postings. Id. Inevitably, disagreements arose among the users, and a lawsuit followed seeking to hold CompuServe liable for a posting on its system.

The district court sketched two paths for determining CompuServe‘s liability. Perhaps the company could be considered a “publisher,” someone strictly liable for repeating defamatory statements no matter the company‘s knowledge of what was said and why it might be actionable. Id. at 139. Or the company might be a “distributor,” like “news vendors, book stores, and libraries,” and liable only if the company knew or had reason to know the statement was defamatory. Id. The district court decided CompuServe‘s forum was “in essence an electronic, for-profit library,” with the company having “little or no editorial control over [the forum‘s] contents.” Id. at 140. And because it was merely a distributor, liability could only attach if CompuServe knew the post was defamatory (which it did not). Id. at 140–41.

CompuServe both won praise and stoked worry because the opinion turned on the amount and kind of editorial control exercised by the internet forum, a test that could vary in application from service to service. See, e.g., Jonathan M. Moses & Michael W. Miller, CompuServe Is not Liable for Contents, Wall St. J. (Oct. 31, 1991). Prodigy, for example, sold subscribers on the rigor of its screening and the promise that families could enjoy online entertainment without offensive messages. That suggested Prodigy could be subject to strict liability because it was “the only major commercial [bulletin board] operator that monitor[ed] all public messages by screening them before they [were] posted.” David J. Conner, Note, Cubby v. CompuServe, Defamation Law on the Electronic Frontier, 2 Geo. Mason Indep. L. Rev. 227, 240 (1993).

These predictions proved prescient. Three years later, in Stratton Oakmont, Inc. v. Prodigy Services Company, No. 31063/94, 1995 WL 323710 (N.Y. Sup. Ct. May 24, 1995), Prodigy was sued for hosting allegedly defamatory statements posted on one of its electronic bulletin boards. Id. at *1. Following the reasoning of CompuServe, the Stratton Oakmont court found Prodigy “exercised sufficient editorial control over its computer bulletin boards to render it a publisher with the same responsibilities as a newspaper.” Id. at *3. That meant Prodigy was liable for any defamatory statements on its service. Id. at *3–5. Though it was a non-precedential opinion issued by a state trial court judge, Stratton Oakmont received significant attention, much of it negative.¹⁵ If Stratton Oakmont‘s reasoning stood, online service providers acting to exclude offensive and obscene content would now risk liability for the rest of the material they hosted. See Adam Candeub, Bargaining For Free Speech: Common Carriage, Network Neutrality, and Section 230, 22 Yale J.L. & Tech. 391, 421 (2020).

B.

1. Congress responded vigorously, and a mere nine months after Stratton Oakmont, the President signed the Communications Decency Act of 1996 (CDA) into law as part of the Telecommunications Act of 1996, Pub. L. No. 104–104, 110 Stat. 56. A last-minute addition to the Telecommunications Act, the CDA was initially designed to regulate internet pornography and protect children from obscene and harmful material. See Robert Cannon, The Legislative History of Senator Exon‘s Communications Decency Act: Regulating Barbarians on the Information Superhighway, 49 Fed. Comm. L.J. 51, 52–58 (1996). But a month after Stratton Oakmont, lawmakers introduced the CDA amendment that ultimately became § 230. See Internet Freedom and Family Empowerment Act, H.R. 1978, 104th Cong. (1995). Unlike other aspects of the CDA, § 230‘s “proposal and passage flew under the radar” and “received virtually no opposition or media coverage.” Jeff Kosseff, The Twenty-Six Words That Created The Internet 3 (2019).

As enacted, § 230 created two complementary protections. Section 230(c)(1) directs that “[n]o 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.” And § 230(c)(2)(A) states that “[n]o provider or user of an interactive computer service shall be held liable on account of . . . any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable.” The statute expressly preempts any “cause of action” or “liability” “under any State or local law that is inconsistent with” those provisions. § 230(e)(3).

It is conventional wisdom that § 230 was passed to, at least in part, overrule Stratton Oakmont,¹⁶ a goal that fit within the purpose of the CDA‘s statutory scheme. Most of the CDA‘s provisions sought to protect minors from offensive online material. See Force v. Facebook, Inc., 934 F.3d 53, 78–80 (2d Cir. 2019) (Katzmann, J., concurring in part & dissenting in part). But Stratton Oakmont‘s reasoning undercut incentives for computer services to limit access to offensive material. After all, it was precisely Prodigy‘s attempt to moderate its platform to provide a family-friendly environment that led to vast tort liability. See Stratton Oakmont, 1995 WL 323710, at *1–5. By overruling Stratton Oakmont, Congress encouraged private action to complement the CDA‘s regulations and bolster efforts to reduce the spread of indecent material on the internet. See Zeran v. Am. Online, Inc., 129 F.3d 327, 331 (4th Cir. 1997).

2. But from the very start, courts held § 230 did much more than overrule Stratton Oakmont‘s publisher-liability theory. And they almost all followed Zeran v. America Online, Inc., which read § 230(c)(1) to immunize an interactive computer service provider‘s “exercise of a publisher‘s traditional editorial functions—such as deciding whether to publish, withdraw, postpone or alter content.” 129 F.3d at 330. This broad immunity was broadened even further when Zeran held that § 230(c)(1) barred both publisher and distributor liability. Id. at 331–34. Though Zeran has been criticized as inconsistent with the text, context, and purpose of § 230 (and was decided in an era where those traditional tools of construction were rarely consulted), the opinion was cut-and-paste copied by courts across the country in the first few years after the statute arrived. See Malwarebytes, Inc. v. Enigma Software Grp. USA, LLC, 141 S. Ct. 13, 15–18 (2020) (Thomas, J., statement respecting denial of certiorari); Calise v. Meta Platforms, Inc., 103 F.4th 732, 746–47 (9th Cir. 2024) (Nelson, J., concurring); Candeub, Bargaining For Free Speech, supra, at 423–28.

Today, § 230 rides in to rescue corporations from virtually any claim loosely related to content posted by a third party, no matter the cause of action and whatever the provider‘s actions. See, e.g., Gonzalez v. Google LLC, 2 F.4th 871, 892–98 (9th Cir. 2021), vacated, 598 U.S. 617 (2023); Force, 934 F.3d at 65–71. The result is a § 230 that immunizes platforms from the consequences of their own conduct and permits platforms to ignore the ordinary obligation that most businesses have to take reasonable steps to prevent their services from causing devastating harm.

C.

But this conception of § 230 immunity departs from the best ordinary meaning of the text and ignores the context of congressional action. Section 230 was passed to address an old problem arising in a then-unique context, not to “create a lawless no-man‘s-land” of legal liability. Fair Hous. Council of San Fernando Valley v. Roommates.com, LLC, 521 F.3d 1157, 1164 (9th Cir. 2008) (en banc).

As with all cases involving the interpretation of statutes, our job in interpreting § 230‘s text is to “give effect to the legislature‘s charge,” “stated through the ‘ordinary meaning . . . at the time Congress enacted the statute.‘” OI Eur. Grp. B.V., 73 F.4th at 165 (quoting Perrin v. United States, 444 U.S. 37, 42 (1979)). Courts must take care to construe a statute‘s terms in light of “background understandings and the structure and circumstances of the [legislative] utterance.” Herrmann v. Cencom Cable Assocs., Inc., 978 F.2d 978, 982 (7th Cir. 1992). A task that necessarily includes consideration of the legal “backdrop against which Congress” acted. Stewart v. Dutra Constr. Co., 543 U.S. 481, 487 (2005); see also Biden v. Nebraska, 143 S. Ct. 2355, 2378 (2023) (Barrett, J., concurring) (“[T]extualists, like all interpreters,” read “text in context. . . . Context is not found exclusively within the four corners of a statute. Background legal conventions, for instance, are part of the statute‘s context.” (cleaned up)).¹⁷

1. Section 230(c)(1) directs that TikTok not be “treated as the publisher . . . of any information provided by another information content provider.”¹⁸ Congress enacted § 230 mindful of the recent and widely discussed online service provider tort cases drawing the publisher-distributor distinction, as well as decades of state and federal law apportioning liability for electronic transmissions along the same line. That points to the best reading of § 230(c)(1) as adopting the meaning of “publisher” used by Stratton Oakmont and CompuServe. See George v. McDonough, 596 U.S. 740, 746 (2022) (“Where Congress employs a term of art obviously transplanted from another legal source, it brings the old soil with it.” (cleaned up)). So when § 230(c)(1) prohibits “treat[ing]” TikTok as the “publisher” of videos posted by third parties, that means TikTok cannot be liable for the mere act of hosting those videos. See Malwarebytes, 141 S. Ct. at 14–16 (Thomas, J., statement respecting denial of certiorari); Doe ex rel. Roe v. Snap, Inc., 88 F.4th 1069, 1070–72 (5th Cir. 2023) (Elrod, J., dissenting from denial of rehearing en banc);

Candeub, Reading Section 230 as Written, supra, at 146–51. It cannot, in short, be held liable as a publisher.

But § 230(c)(1) does not immunize more. It allows suits to proceed if the allegedly wrongful conduct is not based on the mere hosting of third-party content, but on the acts or omissions of the provider of the interactive computer service. This is where Zeran went astray, wrongly reasoning that distributor liability “is merely a subset, or a species, of publisher liability.” 129 F.3d at 332. It is true that “[s]ources sometimes use language that arguably blurs the distinction between publishers and distributors.” Malwarebytes, 141 S. Ct. at 15 (Thomas, J., statement respecting denial of certiorari). But understanding § 230(c)(1)‘s use of “publisher” to subsume distributor liability conflicts with the context surrounding § 230‘s enactment. Both CompuServe and Stratton Oakmont saw two distinct concepts. See CompuServe, 776 F. Supp. at 138–41; Stratton Oakmont, 1995 WL 323710, at *1–5. So did the common law of common carriers. It is implausible to conclude Congress decided to silently jettison both past and present to coin a new meaning of “publisher” in § 230(c)(1). See Malwarebytes, 141 S. Ct. at 14–16 (Thomas, J., statement respecting denial of certiorari); Doe v. Am. Online, Inc., 783 So. 2d 1010, 1023–25 (Fla. 2001) (Lewis, J., dissenting).

2. Properly read, § 230(c)(1) says nothing about a provider‘s own conduct beyond mere hosting.¹⁹ A conclusion confirmed by § 230(c)(2), which enumerates acts that platforms can take without worrying about liability.²⁰ See David L. Shapiro, Continuity and Change in Statutory Interpretation, 67 N.Y.U. L. Rev. 921, 942 (1992) (“[A]ll legislation occurs against a background of customs and understandings of the way things are done. . . . [A] speaker who is issuing an order or prohibition is likely to focus on what is being changed and to expect the listener to understand that, so far as this communication is concerned, all else remains the same.“).²¹

3. What does all this mean for Anderson‘s claims? Well, § 230(c)(1)‘s preemption of traditional publisher liability precludes Anderson from holding TikTok liable for the Blackout Challenge videos’ mere presence on TikTok‘s platform. A conclusion Anderson‘s counsel all but concedes. But § 230(c)(1) does not preempt distributor liability, so Anderson‘s claims seeking to hold TikTok liable for continuing to host the Blackout Challenge videos knowing they were causing the death of children can proceed. So too for her claims seeking to hold TikTok liable for its targeted recommendations of videos it knew were harmful. That is TikTok‘s own conduct, a subject outside of § 230(c)(1). Whether that conduct is actionable under state law is another question. But § 230 does not preempt liability on those bases.²²

* * *

“It used to be said that there were three great influences on a child: home, school, and church. Today, there is a fourth great influence . . . .” Newton N. Minow, Speech Before the Nat‘l Ass‘n of Broads. (May 9, 1961), reprinted in Newton N. Minow, Television and the Public Interest, 55 Fed. Comm. L.J. 395, 399 (2003). When Commissioner Minow spoke of the perils and promise of television, the internet was still two decades from its earliest form. But his description of a “procession of game shows, . . . formula comedies about totally unbelievable families, blood and thunder, mayhem, violence, sadism, murder, . . . more violence, and cartoons” captures the dreary state of the modern internet. Id. at 398. The marketplace of ideas, such as it now is, may reward TikTok‘s pursuit of profit above all other values. The company may decide to curate the content it serves up to children to emphasize the lowest virtues, the basest tastes. It may decline to use a common good to advance the common good.

But it cannot claim immunity that Congress did not provide. For these reasons, I would affirm the District Court‘s judgment as it relates to any of Anderson‘s claims that seek to hold TikTok liable for the Blackout Challenge videos’ mere existence on TikTok‘s platform. But I would reverse the District Court‘s judgment as it relates to any of Anderson‘s claims that seek to hold TikTok liable for its knowing distribution and targeted recommendation of the Blackout Challenge videos. Accordingly, I concur in the judgment in part and dissent in part.

Notes

We draw the facts from the complaint, accept them as true, , and “view[] them in the light most favorable to [the] plaintiff,” (citation omitted). “[W]e disregard legal conclusions and ‘recitals of the elements of a cause of action, supported by mere conclusory statements.‘” (quoting ). Assumptions that find no support in the First Amendment, which “was not designed or originally understood to provide a font of judicially crafted doctrines protecting expressive freedom.” Jud Campbell, Natural Rights and the First Amendment, 127 Yale L.J. 246, 320 (2017). And “[t]he problem with Section 230 is that in a bout of free speech zeal, courts have interpreted the law to be far more extensive than it is written or should be.” Daniel Solove, Restoring the CDA Section 230 to What It Actually Says, TeachPrivacy (Feb. 4, 2021), https://teachprivacy.com/restoring-the-cda-section-230-to-what-it-actually-says/.

An algorithm is a set of digital instructions that perform a task. See (citation omitted). Saint Augustine of Hippo, The Confessions of Saint Augustine 42 (Hackett Publishing Co. 2006).

Anderson also brings claims for wrongful death and under Pennsylvania‘s Survival Act, 42 Pa. Cons. Stat. § 8302. Because those claims are derivative of her tort claims, her ability to pursue them depends on whether her tort claims survive the motion to dismiss. See (survival action); (wrongful death). Anderson abandoned her claims under the Pennsylvania Unfair Trade Practices and Consumer Protection Law, 73 Pa. Stat. and Cons. Stat. § 201-1 et seq., and the California Consumer Legal Remedies Act, Cal. Civ. Code § 1750 et seq. See . We must take the well-pleaded factual allegations drawn from the complaint as true. .

Anderson does not challenge the District Court‘s order denying her motion for leave to file an amended complaint. The For You Page displays a unique series of videos to each user based on TikTok‘s algorithm, which “selects which videos are shown to each user based on the user‘s demographics, including age, [and] user interactions such as the videos viewed and shared.” App. 28 (emphasis omitted).

The District Court had jurisdiction pursuant to 28 U.S.C. § 1332(a). We have jurisdiction pursuant to 28 U.S.C. § 1291. We exercise plenary review over a district court‘s order granting a motion to dismiss. (citation omitted). “To survive a motion to dismiss, a complaint must ... plead[] factual content that allows the court to draw the reasonable inference that the defendant is liable for the misconduct alleged.” (citation omitted). See Matt Stoller, Goliath 5–7 (2019).

TikTok is an “interactive computer service,” which is defined as “any information service, system, or access software provider that provides or enables computer access by multiple users to a computer server, including specifically a service or system that provides access to the Internet[.]” 47 U.S.C. § 230(f)(2). See Peter Huber, Law and Disorder in Cyberspace 26 (1997) (“Federal authorities had already been regulating railroads for decades. Congress figured that regulating phones would be much the same.“); see also James B. Speta, A Common Carrier Approach to Internet Interconnection, 54 Fed. Comm. L.J. 225, 261–68 (2002); (Thomas, J., concurring); . Common carriage arrangements pursued a regulatory bargain, with carriers receiving benefits (like immunities from suit and market control) in exchange for increased delivery obligations. See Adam Candeub, Bargaining For Free Speech: Common Carriage, Network Neutrality, and Section 230, 22 Yale J.L. & Tech. 391, 398–413 (2020).

Specifically, “Congress enacted the CDA in response to a state-court decision which held that the provider of an online messaging board could be liable for defamatory statements posted by third-party users of the board.” (citations omitted). Liability for telephone companies came to function much the same way. See Candeub, Reading Section 230 as Written, supra, at 146 n.26.

“The term ‘information content provider’ means any person or entity that is responsible, in whole or in part, for the creation or development of information provided through the Internet or any other [ICS].” 47 U.S.C. § 230(f)(3). It also conformed with the regulation of other common carriers. A railroad, for example, was generally not liable for a passenger‘s unlawful acts facilitated by the train unless the operator knew its service was being used for an unlawful purpose. See Bruce Wyman, Illegality as an Excuse for Refusal of Public Service, 23 Harv. L. Rev. 577, 584 (1910). So too with telegraphs and telephones that had only a duty to “refuse to transmit messages which would implicate [the company] in illegality,” such as communications these companies knew were libelous, obscene, fraudulent, or otherwise used to further some harmful act proscribed by law. See id. at 584–85, 587.

The immunity stems from the statutory language providing, with limited exceptions that do not apply here, that “[n]o cause of action may be brought and no liability may be imposed under any State or local law that is inconsistent” with § 230(c)(1). 47 U.S.C. § 230(e). See Ryan Gerdes, Scaling Back 230 Immunity: Why the Communications Decency Act Should Take a Page from the Digital Millennium Copyright Act‘s Service Provider Immunity Playbook, 60 Drake L. Rev. 653, 656 (2012).

In , the Supreme Court considered social media platforms’ algorithms that construct feeds to relay content to users. . The Court described the platforms at issue in as ones that “cull and organize uploaded posts in a variety of ways. A user does not see everything ... . The platforms will have removed some content entirely; ranked or otherwise prioritized what remains; and sometimes added warnings or labels.” . The Court explained that, by engaging in such activity, the platforms “shape other parties’ expression into their own curated speech products.” . Although “[t]he selection and ranking is most often based on a user‘s expressed interests and past activities,” the Court noted that “it may also be based on more general features of the communication or its creator[,]” particularly given that some platforms have guidelines that “detail the messages and videos that the platform[] disfavor[s.]” . In holding that “expressive activity includes presenting a curated compilation of speech originally created by others[,]” , the Court declined to address “algorithms [that] respond solely to how users act online[,]” . Accordingly, the presence or absence of a platform‘s standards or preferences that govern an algorithm‘s choices may dictate whether the algorithm is expressive speech, (Barrett, J., concurring), as might whether the platform is a “passive receptacle[] of third-party speech ... that emit[s] what [it is] fed” or whether it only responds to specific user inquiries, (Alito, J., concurring in the judgment). See also (Barrett, J., concurring) (distinguishing types of algorithms); (Alito, J., concurring in the judgment) (same). Take the telegraph operator transmitting the statement “John is a crook.” If the operator does not know that John is scrupulously law-abiding, the context necessary to make the statement false and libelous is absent.

We recognize that TikTok‘s first-party speech captures certain third-party speech. However, “exercis[ing] ‘editorial discretion in the selection and presentation’ of content” qualifies as “‘speech activity’ ... [whether] the content comes from third parties [or] it does not.” (first alteration in original) (quoting ). Lawrence Lessig, The Future of Ideas 147 (2001).

We reach this conclusion specifically because TikTok‘s promotion of a Blackout Challenge video on Nylah‘s FYP was not contingent upon any specific user input. Had Nylah viewed a Blackout Challenge video through TikTok‘s search function, rather than through her FYP, then TikTok may be viewed more like a repository of third-party content than an affirmative promoter of such content. Given the type of algorithm alleged here, we need not address whether § 230 immunizes any information that may be communicated by the results of a user‘s search of a platform‘s content. We need not address in this case the publisher/distributor distinction our colleague describes, nor do we need to decide whether the word “publisher” as used in § 230 is limited to the act of allowing third-party content to be posted on a website an ICS hosts, as compared to third-party content an ICS promotes or distributes through some additional action, because, in this case, the only distribution at issue is that which occurred via TikTok‘s algorithm, which as explained herein, is not immunized by § 230 because the algorithm is TikTok‘s own expressive activity. Id. at 148.

We recognize that this holding may be in tension with , where we held that § 230 immunized an ICS from any liability for the platform‘s failure to prevent certain users from “transmit[ing] harmful online messages” to other users. . We reached this conclusion on the grounds that § 230 “bar[red] ‘lawsuits seeking to hold a service provider liable for ... deciding whether to publish, withdraw, postpone, or alter content.‘” (quoting ). , however, did not involve an ICS‘s content recommendations via an algorithm and pre-dated . Similarly, our holding may depart from the pre- views of other circuits. See, e.g., (“[R]ecommendations and notifications are not content in and of themselves.“); (“Merely arranging and displaying others’ content to users through [] algorithms—even if the content is not actively sought by those users is not enough to hold [a defendant platform] responsible as the developer or creator of that content.” (internal quotation marks and citation omitted)); (concluding that § 230 immunity applied because the structure and operation of the website, notwithstanding that it effectively aided sex traffickers, reflected editorial choices related to traditional publisher functions); (adopting by noting that “traditional editorial functions” are immunized by § 230); (immunizing a platform‘s “decision whether to print or retract a given piece of content“); (adopting ); (rejecting an argument that § 230 immunity was defeated where the allegations went to the platform‘s traditional editorial functions). Jeff Kosseff, The Twenty-Six Words That Created The Internet 37 (2019).

To the extent that Anderson still pursues any claims not premised upon TikTok‘s algorithm, we leave to the District Court to determine, among other things, whether, consistent with this Opinion, those claims are barred by § 230. See Appellant‘s Br. at 21 (acknowledging that TikTok‘s “initial action in publishing the Blackout Challenge generally on the TikTok app may very well fall within the protections of the CDA“); Reply Br. at 9 n.1 (acknowledging that certain allegations in Anderson‘s complaint may be barred by the CDA). Id. at 42.

See Kosseff, supra, at 55–71; R. Hayes Johnson Jr., Defamation in Cyberspace: A Court Takes a Wrong Turn on the Information Superhighway in , 49 Ark. L. Rev. 589, 594 & n.10 (1996); Douglas B. Luftman, Note, Defamation Liability for On-Line Services: The Sky Is Not Falling, 65 Geo. Wash. L. Rev. 1071, 1072 (1997) (describing the “apocalyptic reactions in the legal and technical communities“).

See Kosseff, supra, at 48–82; Candeub, Bargaining For Free Speech, supra, at 419–21; (Katzmann, J., concurring in part & dissenting in part). Contemporary commentators, see, e.g., Cannon, supra, at 61–63, 68, early courts, see, e.g., , and even modern proponents of broad § 230 immunity, see, e.g., Tr. Oral Argument at 126, (No. 21-1333) (Counsel for Google: “[O]ne lawsuit freaked out the Congress . . . .“), all agree.

A principle of interpretation with deep roots in the classical legal tradition. Blackstone understood the interpretation of statutes that sought to change the legal status quo to necessarily include consideration of 1) how the law “stood at the making of the act“; 2) “what the mischief was, for which the [then-existing] law did not provide“; 3) “and what remedy the [legislature] hath provided to cure this mischief.” 1 Blackstone, Commentaries *87. All contextual clues aiding the interpretation of the words the legislature enacted. See .

The reference to “speaker” in § 230(c)(1) does not change the meaning of the text. When § 230 was enacted, courts often referred to traditional publisher liability as treating the disseminator of a statement as the “original speaker” subject to the same strict liability. See Jonathan A. Friedman & Francis M. Buono, Limiting Tort Liability for Online Third-party Content Under Section 230 of the Communications Act, 52 Fed. Comm. L.J. 647, 650 (2000). Consistent with common law tort theory, I refer to such claims as publisher liability rather than “third” versus “first-party speech.” Doing so also avoids the confusing commingling of statutory and constitutional language that can conflate the distinct legal meanings of “speech.”

See (Thomas, J., statement respecting denial of certiorari) (“It is hard to see why the protection § 230(c)(1) grants publishers against being held strictly liable for third parties’ content should protect Facebook from liability for its own ‘acts and omissions.‘” (emphasis in original)); cf. (Tymkovich, J., concurring) (“Section 230 only immunizes publishers or speakers for the content of the information from other providers that they make public. The CDA says nothing about immunizing publishers or speakers for their own conduct . . . .” (emphasis in original) (citation omitted)).

See § 230(c)(2) (“No provider or user of an interactive computer service shall be held liable on account of (A) any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable; or (B) any action taken to enable or make available to information content providers or others the technical means to restrict access to material described in [§ 230(c)(2)(A)].“).

Invoking § 230(f)(4)(C)‘s definition of “access software provider,” TikTok argues for a broader reading of § 230(c)(1) to include immunity for any actions taken to “organize” or “reorganize” content. In TikTok‘s view, its targeted recommendations just organize the hosted content. But I do not read a definitional provision defining a different statutory term to expand the scope of § 230(c)(1)‘s “publisher” immunity. Section 230(f)(4)(C), on its own, provides no immunity. It only states that a provider or user of an interactive computer service does not become an “information content provider“—and thereby fall outside the scope of § 230(c)(1)—just by organizing or reorganizing third-party content. One cannot conclude from § 230(f)(4) that because some providers or users of interactive computer services organize information, § 230(c)(1) necessarily immunizes that conduct. Section 230(f)(4) just loops the reader back to § 230(c)(1) to determine the meaning of “treat[] as the publisher . . . of any information provided by another information content provider.”

A word on , a two-decade-old decision that decided very little. involved a disgruntled former subscriber to AOL‘s chat room service who filed a pro se complaint that was “not especially clear.” . By the time his case made it to this Court, Green‘s main complaint seems to have been that AOL “negligently failed to live up to its contractual obligations” by failing to kick certain third-party users off AOL‘s platform after they sent Green a virus through AOL and posted defamatory statements about him in a chat room. See ; Brief for Appellant, (No. 01-1120), 2002 WL 32397368, at *1–2, *4, *13–14. We explained that “[t]he only question” presented on appeal was “whether holding AOL liable for its alleged negligent failure to properly police its network for content transmitted by its users” was barred by § 230(c)(1). . In a single, three-sentence paragraph of analysis, we answered that question in the affirmative, holding that Green‘s claims were barred by § 230(c)(1) because they sought “to hold AOL liable for . . . actions quintessentially related to a publisher‘s role.” . Exactly what “failure to properly police its network” meant is also “not especially clear.” But in my view, it is best understood to refer to a provider of an interactive computer service failing to pre-screen third-party content before circulation and failing to actively monitor its service for allegedly harmful content. See, e.g., (describing Green‘s complaint that AOL “did nothing to stop” the initial posting of additional defamatory statements); Brief for Appellee, (No. 01-1120), 2002 WL 32397367, at *8 (explaining that Green‘s complaint did not allege that he “suffer[ed] any damages at any time after” he notified AOL of the third-party information). In other words, all held was that § 230 precluded publisher liability as that term was understood by and . said nothing about whether § 230 immunizes providers or users of interactive computer services for failing to take down harmful content once they receive notice of its presence on the platform (distributor liability).

Read the detailed case summary

OPINION OF THE COURT

I

A1

B

II5

III

I.

A.

B.

II.

A.

B.

C.

Notes

A¹

II⁵