The Digital Frontier on Trial: School Districts Versus Social Media Giants

The American classroom has become a primary battleground for one of the most consequential shifts in modern products liability law. As of early 2026, more than 1,000 school districts have filed lawsuits against social media platforms including Meta Platforms, Inc. (Facebook and Instagram), Google, LLC and Alphabet Inc. (YouTube), Snap Inc. (Snapchat), and Bytedance Inc. and TikTok Inc. (TikTok). This includes 45 of Florida’s 67 school districts, which have filed their own lawsuits. These co-ordinated actions represent a doctrinal evolution: the movement from traditional claims involving physical defects in tangible goods to allegations of psychological and neurobiological injuries from an online product’s embedded algorithmic design “defects.”

At issue is a foundational legal question: When does a digital platform cease to be a neutral publisher and become a product manufacturer?

From Playground Disruption to Federal MDL

For more than a decade, school administrators have documented a steep rise in student anxiety, depression, self-harm, and behavioural disruption. In fact, as of 2017, Miami-Dade and Broward middle school students reported all-time high incidences of suicidal ideation of 17.7% and 24.3%, respectively (Youth Risk Behavior Survey 2017 Results (Middle School), Miami-Dade Cnty. Pub. Schs.; Youth Risk Behavior Survey, Broward Cnty. Pub. Schs). While early explanations focused on broad societal forces, the battle now squarely targets the design architecture of social media platforms. In June 2024, the US Surgeon General called for a health warning label to be placed on social media platforms based on social media’s association “with significant mental health harms for adolescents” (Vivek H. Murthy, Surgeon General: Why I’m Calling for a Warning Label on Social Media Platforms, N.Y. Times (June 17, 2024), see New York Times.) And in 2024, Florida Governor Ron DeSantis signed into law a statute banning children under the age of 14 from creating social media accounts. The law has since been challenged by tech-industry groups on the grounds that it violates First Amendment rights.

In 2023, the federal Judicial Panel on Multidistrict Litigation centralised in a San Francisco federal court numerous personal injury, state attorney general civil enforcement, municipality, and school district cases into one multidistrict litigation (MDL), styled In re Social Media Adolescent Addiction/Personal Injury Products Liability Litigation, MDL No. 3047 (N.D. Cal.), before Judge Yvonne Gonzalez Rogers.

The school districts generally allege that social media companies defectively designed and engineered their products to maximise engagement through compulsive-use mechanisms. The school districts further allege that companies externalised the resulting mental health costs onto public schools, forcing diversion of educational funds to crisis intervention, counselling, behavioural monitoring software, police enforcement, and infrastructure repair tied to viral social media challenges.

The legal framing echoes the US Supreme Court’s longstanding recognition that products liability focuses exclusively on manufacturer conduct. As the Supreme Court observed in Bates v Dow Agrosciences LLC, products liability law “require[s] manufacturers to design reasonably safe products, to use due care in conducting appropriate testing of their products, to market products free of manufacturing defects.” 544 U.S. 431, 444 (2005). See also Air & Liquid Sys. Corp. v DeVries, 586 U.S. 446, 452-53 (2019) (Basic tort-law principles “impose[] ‘a duty to exercise reasonable care’ on those whose conduct presents a risk of harm to others. For the manufacturer of a product, the general duty of care includes a duty to warn when the manufacturer ‘knows or has reason to know’ that its product ‘is or is likely to be dangerous for the use for which it is supplied’ . . . .”) (citations omitted). In the MDL, the school districts ‒ among other plaintiffs ‒ argue that algorithmic amplification systems fall squarely within that tradition.

Section 230 and the “Design Versus Speech” Divide

One of the central defences the social media companies have lodged has been Section 230 of the Communications Decency Act, 47 U.S.C. §230(c)(1), which provides 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.”

Historically, courts interpreted this language broadly. In Zeran v America Online, Inc., for example, the Fourth Circuit Court of Appeals held: “Congress recognized the threat that tort-based lawsuits pose to freedom of speech in the new and burgeoning Internet medium.” 129 F.3d 327, 330 (4th Cir. 1997).

However, many recent court rulings have significantly narrowed the scope of Section 230 immunity where claims target platform design rather than third-party content. In Lemmon v Snap, Inc., the Ninth Circuit Court of Appeals held that Section 230 did not bar claims alleging Snap’s “Speed Filter” was defectively designed, drawing a critical distinction:

“Snap ‘is being sued for the predictable consequences of’ designing Snapchat in such a way that it allegedly encourages dangerous behavior. Section 230 does not shield Snap from liability for such claims.”

995 F.3d 1085, 1094 (9th Cir. 2021) (citation omitted). Other courts followed suit. See, eg, Anderson v TikTok, Inc., 116 F.4th 180, 184 (3d Cir. 2024) (finding no Section 230 immunity when a social media company “curate[s] compilations of others’ content via their expressive algorithms”); Doe v MG Freesites, LTD, 2024 WL 5339485, at *8-*11 (N.D. Ala. Dec. 19, 2024), motion to certify appeal denied sub nom., Doe # 1 v MG Freesites, LTD, 2025 WL 1314179 (N.D. Ala. May 6, 2025); Huckabee v Meta Platforms, Inc., 2024 WL 4817657, at *4-*5 (D. Del. Nov. 18, 2024).

In the MDL proceedings, Judge Gonzalez Rogers rejected blanket Section 230 immunity for claims focused on algorithmic design and engagement features, emphasising “[t]he defects pled as part of such allegations are not equivalent to speaking or publishing [but are manufacturing decisions that] can be fixed by defendants without altering the publishing of third-party content.” In re Soc. Media Adolescent Addiction/Pers. Inj. Prods. Liab. Litig., 702 F. Supp. 3d 809, 829 (N.D. Cal. 2023) (emphasis in original); accord In re Soc. Media Adolescent Addiction/Pers. Inj. Prods. Liab. Litig., 777 F. Supp. 3d 1016, 1022 (N.D. Cal. 2025) (“allegations relating to particular features of defendants’ platforms [are not] barred by Section 230”) (emphasis in original).

Taken together, these rulings reflect a developing judicial consensus: Section 230 does not categorically immunise platform design choices merely because they operate through content display. And, although the Supreme Court has not resolved the outer boundaries of Section 230 immunity, several of the justices signalled during a recent oral argument their concern over extending immunity to algorithmic recommendations by social media companies, see Transcript of Oral Argument at 17, 47, 49, Gonzalez v Google LLC, 598 U.S. 617 (2023) (No. 21-1333), and the Court has also found ‒ albeit in the First Amendment context ‒ that a platform’s algorithm reflects “editorial judgments” about “compiling the third-party speech it wants in the way it wants” and is the platform’s own “expressive product.” Moody v NetChoice, LLC, 603 U.S. 707, 718, 731-32 (2024).

Algorithmic Architecture as Product Design

Many school districts characterise social media features such as infinite scroll, push notifications, variable reward systems, streak mechanisms, and AI-driven recommendations as a “casino that fits in your pocket,” a phrase coined by New York Times reporter Max Fisher in Chaos Machine: The Inside Story of How Social Media Rewired Our Minds and The World (2022). Courts have increasingly accepted that such features can constitute product design elements subject to ordinary negligence or strict liability analysis. As the Ninth Circuit explained in Lemmon, “[t]he duty to design a reasonably safe product is fully independent of Snap’s role in monitoring or publishing third-party content.” 995 F.3d at 1093. This distinction reframes social media platforms as engineered behavioural systems rather than passive conduits.

The analogy to tobacco litigation is not rhetorical. In Kerrivan v R.J. Reynolds Tobacco Co., the court found that:

“The Tobacco Companies knew that cigarettes were addictive and linked to serious health conditions. Despite this knowledge, they not only concealed and denied the adverse health effects of smoking, but they also deliberately designed their product to deliver higher doses of nicotine to smokers, all the while making cigarettes milder and more inhalable, and therefore more addictive.”

953 F.3d 1196, 1209 (11th Cir. 2020); see also United States v Philip Morris USA Inc., 566 F.3d 1095, 1120-21 (D.C. Cir. 2009) (recognising the “volumes of evidence” demonstrating that tobacco company executives “knew about the negative health consequences of smoking, the addictiveness and manipulation of nicotine, [and] the harmfulness of secondhand smoke”). In fact, Meta employees often made the analogy comparing social media and tobacco products in internal company documents. School districts now argue that internal company research ‒ most notably the “Facebook Files,” released by whistleblower Frances Haugen ‒ demonstrates internal company knowledge of youth harm coupled with continued optimisation for engagement metrics (and thus, advertising dollars).

The Breathitt County Bellwether Trial

In June 2026, the first federal school district bellwether case in the MDL, Breathitt County Board of Education v Meta Platforms, Inc., No. 4:22-cv-03047 (N.D. Cal.), is scheduled for trial before Judge Gonzalez Rogers. The school district seeks approximately USD3 million from the defendants. Bellwether trials serve as valuation proxies in MDL practice. As the Manual for Complex Litigation (Fourth) §22.315 (2025) explains, bellwether trials are intended to “produce a sufficient number of representative verdicts and settlements” that provide guidance for the resolution of the remaining cases. Thus, a significant verdict in favour of the school district could produce enormous settlement pressure on the social media platforms.

Parallel Proceedings: New Mexico and California

The New Mexico attorney general trial

On 24 March 2026, in a Santa Fe, New Mexico courtroom, a jury rendered the first verdict in a “social media addiction” lawsuit, New Mexico v Meta Platforms, Inc., No. D-101-CV-2023-02838 (N.M. 1st Jud. Dist. Ct.), ordering Meta to pay USD375 million in a case brought by the New Mexico Attorney General alleging unfair and unconscionable business practices by the social media giant under New Mexico’s Unfair Practices Act (UPA), N.M. Stat. Ann. §57-12-1, et seq. According to the state, Meta made a conscious decision in 2016 to allow its algorithm to control Facebook and Instagram’s feeds to ensure maximum engagement, leading to what Meta termed “problematic use,” but the state countered was actually classic addiction. Meta has vowed to appeal. Of interest, the New Mexico trial saw introduction of content-driven evidence, including that related to child sexual abuse material, ultimately bypassing Section 230 restrictions.

This verdict marked the first jury evaluation of the “algorithmic addiction” claims, awarding the maximum USD5,000 penalty allowable under the UPA per New Mexico youth user under the age of 18. See N.M. Stat. Ann. §57-12-11. The outcome provides important information to attorneys and their clients in these cases about the ability of a jury to conceptualise algorithmic reinforcement systems as analogous to addictive product design ‒ a theory initially recognised in tobacco jurisprudence.

Importantly, this trial only resolved New Mexico’s UPA claims against Meta; its public nuisance theory ‒ based on allegations that Meta’s platform architecture created what the state characterises as a “marketplace for predators” ‒ will be tried separately by the judge, under a bench trial, at a later date. The public nuisance doctrine historically applies to conditions that interfere with public rights (Restatement (Second) of Torts §821(B)), and many courts have held that conduct that interferes with public health can constitute a public nuisance. See City of Huntington, W. Va. v AmerisourceBergen Drug Corp., 157 F.4th 547, 574-75 (4th Cir. 2025); In re JUUL Labs, Inc., Mktg., Sales Pracs., & Prods. Liab. Litig., 497 F. Supp. 3d 552, 647 (N.D. Cal. 2020); In re Nat’l Prescription Opiate Litig., 2021 WL 4952468, at *4 (N.D. Ohio Oct. 25, 2021). Under its theory of public nuisance and remedies available under this claim, New Mexico will likely focus more so on injunctive and equitable relief, including seeking changes to Meta’s algorithms and Meta’s payment of funds for programs to abate existing and potential harms to its youth users. See, eg, City of Huntington, 157 F.4th at 574 (“conclud[ing] that West Virginia law permits abatement of a public nuisance to include a requirement that a defendant pay money to fund efforts to eliminate the resulting harm to the public”); In re Peabody Energy Corp., 958 F.3d 717, 724 (8th Cir. 2020) (“In California, a party who commits a public nuisance can be ordered to pay into a fund, overseen by a receiver, to remedy or eliminate the hazard complained of rather than being ordered to clean up the nuisance themselves.”).

The Los Angeles, California P.F. trial

Just 24 hours after the verdict in New Mexico, and on the ninth day of deliberations, a jury in a Los Angeles, California state court rendered a verdict in a bellwether personal injury trial, P.F. v Meta Platforms, Inc., No. 23SMCV03371 (Cal. Super. Ct., Los Angeles Cnty.), involving a young woman alleging severe psychological harm from early Instagram and YouTube exposure. The jury awarded the plaintiff USD3 million in compensatory damages, with another USD3 million in punitive damages, finding that Meta and Google acted with malice, fraud, or oppression. Although this case centred around personal injury harms and defences, it served as a one-two punch against two of the world’s largest, most powerful technology companies and a stark warning that its arguments that its social media platforms are social “goods” and “not addictive” are falling flat.

Impact on Florida

The State of Florida is also the epicentre of the litigation. In October 2023, before her appointment to the US Senate by Florida Governor DeSantis, Attorney General Ashley Moody filed a civil enforcement action against Meta under Florida’s consumer protection law, alleging that Meta “knowingly designed and deployed harmful features on Instagram and its other social media platforms that purposefully addict children and teens.” Press Release, Off. of the Att’y Gen., Attorney General Moody Takes Legal Action Against Meta to Protect Children (Oct. 24, 2023), see My Florida Legal. Following already established precedent from around the country, a Florida state court trial judge denied Meta’s motion to dismiss on both Section 230 and First Amendment grounds. Meta has appealed that ruling. In the meantime, the 45 Florida school district lawsuits are largely pending in the MDL before Judge Gonzalez Rogers. These suits, like the Breathitt case scheduled as the first bellwether trial, seek economic damages to recover the costs the districts have incurred due to the mental health crises they allege are directly attributable to the social media companies’ platforms. Robbins Geller represents both the Miami-Dade County School District and the Broward County School District, the two largest in Florida and the third and sixth largest in the United States, respectively.

Tobacco Moment for Big Tech?

The parallels to 1990s tobacco litigation are increasingly difficult to ignore. In R.J. Reynolds and Philip Morris, the courts concluded that decades of internal research and external denial established systemic corporate misconduct. The US Surgeon General likened tobacco to social media companies when calling for social media platform warning labels in June 2024, citing the success of Surgeon General’s warning labels on cigarette packs introduced by Congress in the 1960s. He went on to state that “[social media] harms are not a failure of willpower and parenting; they are the consequence of unleashing powerful technology without adequate safety measures, transparency[,] or accountability.” See Murthy above. If juries find that social media companies knowingly engineered compulsive engagement systems harmful to minors, the financial exposure could be extraordinary.

A plaintiff verdict in Breathitt County (on top of New Mexico’s and P.F.’s recent victories) would not merely resolve one school district’s claims ‒ it could establish a litigation template for thousands of public entities nationwide and potentially catalyse a global settlement fund for youth mental health initiatives.

For the first time, courts are confronting whether the architecture of the digital world itself can be a legally cognisable product defect. The outcome may redefine not only Section 230 jurisprudence, but the contours of modern products liability law in the algorithmic age.