Amicus Brief in Lawsuit Against Meta Over Buffalo Supermarket Shooting and Allegedly Addictive Design

From Jane Bambauer’s and my brief in Patterson v. Meta Platforms, Inc., pending now before New York’s highest court (for more on the case, see the lower court opinions)—many thanks to our excellent local counsel Vivek Tata, Molly O’Keefe, and Alastair Pearson (Elsberg, Baker & Maruri PLLC) for all their help with the brief:

[* * *]

Plaintiffs-Appellants are suing Defendants-Respondents because Defendants-Respondents hosted and displayed to users what appears to be constitutionally protected third-party speech. Such a theory of liability is inconsistent with the First Amendment, as many cases rejecting negligence liability based on the communicative impact of speech have held.

This First Amendment protection cannot be avoided on the grounds that defendants’ speech products are supposedly “addictive.” There is no “addictive speech” exception to the First Amendment. The government cannot interfere with speech producers’ rights to choose how to format and present speech, on the theory that the products are “addictive.” And it cannot interfere with viewers’ access to speech products that give the viewers more of what they want, on the theory that the viewers should not want it or are somehow being manipulated into wanting it.

Indeed, using “engagement-maximizing” techniques is a commonplace, and constitutionally protected, feature of a great deal of speech: novels, TV series, political speeches, news programming, and more. By way of analogy, the Free Exercise Clause would preclude the government from restricting religious practices that supposedly psychologically “addict” congregants through such “engagement-maximizing” tools that supposedly “force-feed[]” religious material to congregants (see infra Part II). Likewise, the Free Speech Clause and Free Press Clause preclude the government from using similar arguments to restrict the design of speech products.

This First Amendment protection also cannot be avoided on the grounds that defendants’ speech products use AI algorithms. The First Amendment protects listener rights as well as speaker rights. And listeners have the right to obtain ideas and information without government interference, whether the ideas and information are curated exclusively by humans or with the help of computerized algorithms.

[I.] Civil Liability Generally Cannot Be Premised on Constitutionally Protected Speech That Falls Outside Narrow Exceptions (Such as for Incitement)

The Plaintiffs do not allege that the speech that Payton Gendron viewed in this case consisted of constitutionally unprotected incitement, or fell within any other First Amendment exception. “The government may not prohibit speech because it increases the chance an unlawful act will be committed ‘at some indefinite future time.’” Ashcroft v. Free Speech Coal., 535 U.S. 234, 253 (2002) (quoting Hess v. Indiana, 414 U.S. 105, 108 (1973) (per curiam)). The incitement exception to the First Amendment is narrowly limited to speech that “is directed to inciting or producing imminent lawless action and is likely to incite or produce such action.” Brandenburg v. Ohio, 395 U.S. 444, 447 (1969) (per curiam). That is also true of hateful speech and of speech that is aimed at “radicalizing” people, as in Brandenburg itself. And if speech is protected by the First Amendment, an online platform that provides access to a selection of that speech is also protected. See Moody v. NetChoice, LLC, 603 U.S. 707, 728 (2024).

The First Amendment also generally restricts civil liability for such speech, including under negligence and related torts, and not just criminal punishment. See, e.g., the cases cited in the joint brief for Defendants-Respondents Alphabet Inc., Google LLC, YouTube LLC, Meta Platforms, Inc., Instagram LLC, and Reddit, Inc. at 52 & n.20. And that extends to Plaintiffs’ design defect theory, which is a “‘negligence-inspired’ concept,” Patterson Br. at 30, given that “design defect” and “negligent design” cases generally “amount[] to the same thing.” Reis v. Volvo Cars of N. Am., 24 N.Y.3d 35, 42 (2014); see also Denny v. Ford Motor Co., 87 N.Y.2d 248, 257–58 (1995) (“The adoption of this risk/utility balance as a component of the ‘defectiveness’ element has brought the inquiry in design defect cases closer to that used in traditional negligence cases ….”). Negligence liability based on the communicative impact of speech can thus be imposed only when the speech falls within one of the “narrowly defined” “recognized exceptions to First Amendment protection,” such as incitement. Yakubowicz v. Paramount Pictures Corp., 404 Mass. 624, 630 (1989).

And this principle applies even when speech reaches underage viewers. Cf. Brown v. Ent. Merchs. Ass’n, 564 U.S. 786, 794–95 (2011) (holding that minors generally have full First Amendment rights, except as to the historically recognized exception of obscene-as-to-minors speech). Thus, for instance, one appellate court has precluded negligence liability for a TV program that led some underage viewers to sexually attack a small child in copying a scene shown on the program. Olivia N. v. NBC, 126 Cal. App. 3d 488, 492–95 (1981).

Another rejected a claim that a TV program was negligent for permitting a dangerous stunt to be broadcast and for failing to warn plaintiffs’ minor child of the dangers of the stunt, on the grounds that the speech did not fall within one of the “classes of speech which may legitimately be proscribed,” which is to say a First Amendment exception. DeFilippo v. NBC, 446 A.2d 1036, 1038, 1040 (R.I. 1982). Courts have likewise rejected liability for negligence even when the speech involved a televised stunt that caused a child who emulated the act to be partially blinded. Walt Disney Prods. v. Shannon, 276 S.E.2d 580, 581 (Ga. 1981).

Protection for speech producers and distributors against negligence liability makes sense. Allowing negligence claims based on otherwise protected speech—speech that does not fall within one of the narrow First Amendment exceptions—”would invariably lead to self-censorship by broadcasters in order to remove any matter that may … lead to a law suit.” DeFilippo, 446 A.2d at 1041. This would in turn violate defendants’ “right to make their own programming decisions” (even when the defendants are broadcasters and thus seen as having a more “limited” First Amendment right than other speakers). Id.

And allowing this sort of liability would also violate “the paramount rights of the viewers to suitable access to ‘social, esthetic, moral, and other ideas and experiences.’” Id. at 1041–42 (citations omitted) (emphasis added). Such negligence liability would “open the Pandora’s Box” and “have a seriously chilling effect on the flow of protected speech through society’s mediums of communication.” Walt Disney, 276 S.E.2d at 583. “Numerous courts have pointed out that any attempt to impose tort liability on persons engaged in the dissemination of protected speech involves too great a risk of seriously chilling all free speech.” Waller v. Osbourne, 763 F. Supp. 1144, 1151 (M.D. Ga. 1991), aff’d mem., 958 F.2d 1084 (11th Cir. 1992).

The cost-benefit balancing at the heart of an ordinary negligence claim or a design defect claim is also too vague and potentially speech-restrictive to be permissible when it comes to regulating constitutionally protected speech. To be sure, even otherwise protected speech might potentially be restrictable if the restriction passes strict scrutiny. But “strict scrutiny” in free speech cases is an “unforgiving” standard, and “is fatal in fact absent truly extraordinary circumstances.” Free Speech Coal., Inc. v. Paxton, 606 U.S. 461, 484–85 (2025).

In particular, “[c]rucial to the safeguard of strict scrutiny” required in First Amendment cases “is that we have a clear limitation, articulated in the legislative statute or an administrative regulation, to evaluate.” James v. Meow Media, Inc., 300 F.3d 683, 697 (6th Cir. 2002). No such clear limitation is present when a factfinder “evaluating [the plaintiff's] claim of negligence would ask whether the defendants took efficient precautions … that would be less expensive than the amount of the loss.” Id. Courts “cannot adequately exercise [their] responsibilities to evaluate regulations of protected speech, even those designed for the protection of children, that are imposed pursuant to a trial for tort liability.” Id.

Perhaps a statute or regulation could constitutionally set forth precise limits on social media platform design aimed at preventing harm to minors. But see Brown, 564 U.S. at 799–803 (striking down a statute that limited minors’ access to violent video games, on the grounds that the state had not adequately proved that the law was indeed necessary to serve a compelling government interest). But a vague reasonable design standard, a cost-benefit balancing test that would be enforced unpredictably by different juries in different cases, is not constitutional. “‘Whither our children’ … is an important question, but their guidance through the regulation of protected speech should be directed in the first instance to the legislative and executive branches of state and federal governments, not the courts.” James, 300 F.3d at 697.

Nor can Plaintiffs’ claim be distinguished from the preceding claims on the grounds that it targets design rather than content. The First Amendment protects choices about how to present content, not just content itself. See infra Part II.

[II.] Constitutionally Protected Behavior Cannot Be Restricted Simply on The Theory That It Is Psychologically “Addictive”

While unprotected categories of speech “may be prohibited without violating the First Amendment,” Ashcroft, 535 U.S. at 245–46, allegedly “addictive” speech is not one of those categories that can be freely regulated. Just as the First Amendment would preclude regulating religious practices on the grounds that they are allegedly psychologically addictive, so it precludes regulating speech product designs on such grounds.

The allegedly “addictive” elements in this case mainly concern the delivery of speech: “autoplay, infinite scroll, notifications, and social validation features including ‘likes’” and “engagement-maximizing algorithms.” Patterson Br. at 9. But a speaker does not lose constitutional protection by providing speech in a way that induces readers to want more speech.

Many, perhaps most, speakers and distributors seek “maximum engagement,” for their speech, and they often do so for “pecuniary” reasons. Id. at 8. Many speakers and distributors, like social media platforms, try to create content “from which [viewers] cannot look away.” Id. Indeed, using psychologically effective tools to keep listeners engaged and likely to pay (or otherwise benefit) the speaker is likely as old as epic poems, plays, novels, and political oratory. Authors of serialized fiction (print or visual) use cliffhangers to keep people coming back to the next episode. Authors create emotional bonds between the reader and entirely fictional characters.

Indeed, it is a professional inadequacy for authors to fail to exploit our irrational ability to emotionally connect with nonexistent people or keep coming back to a story. And the First Amendment forbids distinguishing between “addictive” features in books and social media. “[W]hatever the challenges of applying the Constitution to ever-advancing technology, ‘the basic principles of freedom of speech and the press, like the First Amendment’s command, do not vary’ when a new and different medium for communication appears.” Brown, 564 U.S. at 790 (quoting Joseph Burstyn, Inc. v. Wilson, 343 U.S. 495, 503 (1952)).

The features that Plaintiffs label as addictive appear to fall into three categories: fostering supposedly intrusive urges through reinforcement (as with “autoplay” and “infinite scroll”); pushing for repeated interactions (as with “notifications”); and “social validation” (as with “likes”). See, e.g., Patterson Br. at 9. All of the features “make choices about what third-party speech to display and how to display” it and about how to “organize and prioritize” such speech, thus “shap[ing] other parties’ expression into their own curated speech products”; this makes them “expressive products” that “receive the First Amendment’s protection.” See Moody, 603 U.S. at 716–17.

Some of the features are themselves expressive. Notifications, for example, communicate information about what content has been posted and who is interacting with content. Social validation features like reposts and likes communicate what other users think about the content.

Some features, such as autoplay and infinite scroll, are likely not themselves expressive. But the First Amendment protects decisions about where, when, and how to speak or publish, even when those particular decisions do not in themselves communicate anything. See, e.g., Watchtower Bible & Tract Soc’y of N.Y., Inc. v. Village of Stratton, 536 U.S. 150, 168–69 (2002) (treating choice to distribute material door-to-door as constitutionally protected); McCullen v. Coakley, 573 U.S. 464, 476–77, 497 (2014) (treating choice to speak face-to-face outside abortion clinic as constitutionally protected); City of Lakewood v. Plain Dealer Publ’g Co., 486 U.S. 750, 756 n.5 (1988) (treating choice to distribute speech for money as constitutionally protected); New York Times Co. v. Sullivan, 376 U.S. 254, 266 (1964) (similar). And even if the expressive value that people get from these decisions about how to publish speech is minor, that is immaterial to the First Amendment analysis: The Free Speech Clause protects even relatively trivial speech, not just great ideas. “Even wholly neutral futilities come under the protection of free speech as fully as do Keats’ poems or Donne’s sermons.” United States v. Stevens, 559 U.S. 460, 479–80 (2010) (cleaned up) (quoting Cohen v. California, 403 U.S. 15, 25 (1971)).

Another way of seeing that allegedly “addictive” features do not strip speech products of constitutional protection is by looking at the other half of the First Amendment: religious freedom. The Supreme Court has recognized that “what is good for First Amendment rights of speech must be good for First Amendment rights of religion as well,” Brown, 564 U.S. at 795 n.3, meaning that if speech can be restricted on some theory, religious freedom could be restricted on such a theory as well.

Conversely, if a theory cannot justify restricting religious practice, it cannot justify restricting speech. Indeed, that’s how the Court reasoned in Brown itself: The Court concluded that surely it could not “be made criminal to admit a person under 18 to church, or to give a person under 18 a religious tract, without his parents’ prior consent,” because such a law would be “obviously an infringement upon the religious freedom of young people and those who wish to proselytize young people.” Id. And based on that, the Court concluded that bans on distributing nonreligious speech to minors without the parents’ consent are likewise unconstitutional. Id.

Claims of “addictiveness” can be made about religion with as much (or as little) evidence as about social media. Many religions also try to “maximize user engagement.” Patterson Br. at 2. Cynics often allege that religions do this to increase “corresponding … revenue.” Id.

Religions also try to reach young people, including “teenagers whose frontal cortexes … are underdeveloped.” Id. at 8. Religious practices may rely on “dopamine response,” id., for their effectiveness—dopamine being a key neurotransmitter for many features of human behavior. See, e.g., Paul M. Butler, Patrick McNamara, Jessica Ghofrani & Raymon Durso, Disease-Associated Differences in Religious Cognition in Patients with Parkinson’s Disease, 33 J. Clin. & Exp. Neuropsychol. 917, 924–26 (2011) (finding that Parkinson’s disease patients exhibit diminished religiosity and hypothesizing that the link is due to dopamine loss); Ed Ergenzinger, Faith, God, and Dopamine, WebMD (May 30, 2025), available at https://​perma.cc​/34TC-D4W8 (last accessed July 28, 2026) (pro­viding a first-person account linking religious delusions during bipolar manic episodes to dopamine production).

Religions are also often accused by those who disapprove of them of being “expressly designed to exploit” people’s “neurologic vulnerabilities.” Patterson Br. at 8. They are likewise accused of “ensnar[ing]” potential converts, id. at 32, making their devotees into “a captive audience,” id., and “force-fe[eding]” religious doctrine to parishioners, id. at 10. To be sure, such accusations may well be highly figurative—but they are equally figurative as to social media platforms.

Likewise, many religions include practices that share particular properties that some say make social media features “addictive,” in the sense of being “techniques that foster persistent, intrusive urges” and that “foster compulsion.” Matthew B. Lawrence, Public Health Law’s Digital Frontier: Addictive Design, Section 230, and the Freedom of Speech, 4 J. Free Speech L. 299, 301, 309 (2024). For instance, many religions—including some of the most popular religions—use repetitive prayers, social validation, guilt, concerns about damnation and salvation, and similar mechanisms to develop and reinforce belief and to discourage people from leaving.

These religious practices, like the practices of social media platforms, keep people engaged through “repeated, daily interactions,” id. at 354, such as praying three or five times a day and saying blessings or grace before each meal. The practices could also be described as “contribut[ing] to behavioral addiction through ‘operant conditioning’ techniques such as intermittent reinforcement and variable reward,” id. at 303—techniques such as having members go to church services in which sermons and the likely intensity of the religious experience differ from day to day.

And the techniques often operate through positive feedback. A “like” button on social media, some argue, makes people “kind of addicted to the feedback.” Kyle Langvardt, Regulating Habit-Forming Technology, 88 Fordham L. Rev. 129, 142 (2019) (quoting Hilary Anderson, Social Media Apps Are “Deliberately” Addictive to Users, BBC News (July 3, 2018), available at https://​www.bbc.com​/news​/technology-44640959). If that is so, then one might equally say that the positive reinforcement that formerly lonely people get from a new religious community would make them “kind of addicted to the feedback” as well. A “like” on social media might indeed spark a dopamine rush and pull a user in, causing them to click to see who liked their post and which post it was—but even more so, gestures of friendship or even brotherhood from members of a religious community provide emotional rewards that may pull a person into the community and make it harder for the person to leave.

But surely the Free Exercise Clause would stop the government from imposing liability on the teachings of religions (whether mainstream religions or supposed “cults”) because of the presence of allegedly “addictive” features. And that is true even if a small number of the religion’s adherents engage in criminal violence as a result of having been “radicalized” by these allegedly “addictive” features.

Likewise, the government cannot target speech products on such grounds as well, unless it has extraordinary evidence that the expressive or distribution features will cause significant harm and little countervailing benefit. When behavior is constitutionally protected—whether by the Free Exercise Clause or the Free Speech Clause—government restrictions on the behavior cannot be justified simply on the grounds that they supposedly prevent potential psychological addiction, or that they reduce the chance that some people may be persuaded to act violently.

[III.] The First Amendment Prevents Government Interference with What Readers and Viewers See, Regardless of Whether AI Algorithms Are Used

Platform users have the right to post speech expressing all sorts of viewpoints free of governmental restrictions, and readers have the right to see such viewpoints free of governmental restrictions. This extends even to speech that includes “abstract teaching of the moral propriety or even moral necessity for a resort to force and violence,” Brandenburg, 395 U.S. at 448 (cleaned up)—whether racist violence, violent Communist revolution, violence against the police, anti-abortion violence, animal rights violence, or any other extremist cause. Platforms have the right to display such speech to users, no matter how engaging such speech may be.

And that is true whether the platforms hand-select such speech or use “AI-driven algorithms,” Patterson Br. at 16, 29, 49, 50, to select it. In either situation, “the First Amendment protects the public’s interest in receiving information.” Pac. Gas & Elec. Co. v. Pub. Utils. Comm’n of Cal., 475 U.S. 1, 8 (1986) (plurality opinion). Speech is constitutionally protected “based not only on the role of the First Amendment in fostering individual self-expression but also on its role in affording the public access to discussion, debate, and the dissemination of information and ideas.” First Nat’l Bank of Boston v. Bellotti, 435 U.S. 765, 783 (1978).

Thus, for instance, the Supreme Court struck down a restriction on foreign governments’ mailing foreign propaganda to U.S. recipients, relying on “the addressee’s First Amendment rights” without having to decide whether foreign governments have First Amendment rights of their own. Lamont v. Postmaster Gen., 381 U.S. 301, 305, 307 (1965). Justice Brennan’s concurrence stressed that it is not clear whether the First Amendment protects “political propaganda prepared and printed abroad by or on behalf of a foreign government,” but concluded that the law was unconstitutional because it violated the recipients’ “right to receive” information, regardless of the senders’ rights to speak. Id. at 307–08 (Brennan, J., concurring).

Just as Americans’ right to receive speech exists regardless of whether the speech is produced by a foreign government, so it exists regardless of whether the information is produced with the help of AI algorithms. The government cannot, for instance, ban AI programs that output analyses, manifestos, or sermons that the government views as conveying harmful ideas—racist ideas, anti-government ideas, pro-violence ideas, blasphemous ideas, or anything else. Any such ban would unacceptably interfere with readers’ ability to receive all viewpoints without government interference.

The government likewise cannot impose liability on speech producers and curators based on the allegedly harmful content promoted to viewers using AI algorithms, since such liability would have the unconstitutional purpose and effect of deterring companies from promoting such content. See supra Part I.

Only one United States Supreme Court opinion hints at different First Amendment treatment for material produced with AI assistance: Justice Barrett’s solo concurrence in Moody, which dealt with platforms’ right not to have viewpoint-neutrality obligations imposed on them by state law. In that opinion, Justice Barrett suggested that, “if a platform’s owners hand the [editorial] reins to an AI tool and ask it simply to remove ‘hateful’ content,” then perhaps no “human being with First Amendment rights made an inherently expressive choice not to propound a particular point of view.” Moody, 603 U.S. at 746 (Barrett, J., concurring) (cleaned up). On that basis, she suggested that it might be permissible to limit the platforms’ editorial discretion in such a situation, on the theory that “technology may attenuate the connection between content-moderation actions (e.g., removing posts) and human beings’ constitutionally protected right to decide for themselves the ideas and beliefs deserving of expression, consideration, and adherence.” Id. (cleaned up) (emphases in original).

But Plaintiffs seek to hold platforms liable because they promoted material composed by a human being with First Amendment rights—the author of extremist but constitutionally protected posts—and thus communicated harmful ideas to human readers who have First Amendment rights of their own. The speaker has made an inherently expressive choice to propound a particular point of view. And the platforms gave effect to that choice by providing that viewpoint to more viewers, often based on the platforms’ well-founded belief that the viewers will be interested in the viewpoint.

Under Plaintiffs’ theory, state law would intercede to impose liability on the platforms because they spread such viewpoints, and would thus use the force of law to pressure platforms into not promoting the viewpoints to users. See, e.g., Patterson Br. at 9 (arguing that the social media defendants’ “engagement-maximizing algorithms” help “actively encourage, assist, and facilitate the spread of racist, antisemitic and terrorist propaganda”); id. at 10 (arguing that defendants targeted Gendron with “violent content and connections which … promoted racism, antisemitism, and gun violence”); id. at 8–9, 16 (arguing that the social media platforms risk “radicalization”). That viewpoint-based approach is unconstitutional regardless of the technology the platforms use to decide which material users will see.

Conclusion

Plaintiffs do not allege that the racist and pro-violence speech that helped lead Payton Gendron to commit murder fits within the narrow incitement exception to the First Amendment. That should be the end of the case.

Speech that is outside the First Amendment exceptions cannot lead to civil liability for those who distribute it. That remains true regardless of whether the distribution uses features that Plaintiffs label psychologically “addictive”: The First Amendment has no addictiveness exception, whether for supposedly psychologically addictive religious practices or supposedly addictive presentation of speech. And the First Amendment also protects human viewers from government attempts to control what they see regardless of whether speech is distributed using AI algorithms.

The post Amicus Brief in Lawsuit Against Meta Over Buffalo Supermarket Shooting and Allegedly Addictive Design appeared first on Reason.com.

Source: https://reason.com/volokh/2026/08/07/amicus-brief-in-lawsuit-against-meta-over-buffalo-supermarket-shooting-and-allegedly-addictive-design/