“The arc of the moral universe is long, but it bends toward justice” has been quoted in a range of contexts, some more apt than others.  When it comes to tort liability, the arc clearly bends toward expanded liability over time with periodic contraction, legislative or judicial.  Whether expanded liability is a good thing depends on your perspective, including the side of the litigation “v.” on which you or your clients tend to be. 

One thing that the recent verdicts in two U.S. social media tort liability trials teaches is that U.S. tort law is in an expansion phase, not just for social media platform companies, but more broadly for companies whose direct or indirect consumer relationships have been governed historically by contract.  For the nearly fifty years that consumers have licensed software, downloaded smart phone applications, logged on to social media platforms, or used any technological interface driven by algorithms, their disputes over these services or intangible goods were governed almost exclusively by contract law.^[1]  Contract principles usually allowed the providers of these new technologies and services to avoid the uncertain playing field of product liability litigation.  That is changing. 

We do not address the specific allegations, evidence, or rulings in the two recent social media trials, or recap the decisions over time on the issue of liability for software and software-driven “products.”^[2]  Nor do we discuss the Communications Decency Act or other federal laws more directed to the intersection of the right to free speech and the duties of a platform with regard to content.  Instead, we hope to provide some big picture insights into what is going on and what comes next, which we think reflects continued and highly leveraged efforts to increase tort liability for technology that is rooted in software at the same time as software will increasingly take on tasks historically delivered by people.

Not Really David v. Goliath

While some segments of the press may tout the resources that high market value companies can expend on their legal needs, modern large-scale tort litigation is often about the resources of the plaintiffs’ bar and the impact of market perception on the public companies they sue.  Large litigations do not typically come about organically.  They are usually funded based on research and analysis about how proposed litigation may play out over the course of five to ten years.  They are launched with targeted advertising campaigns and machinery built to gather people or entities to turn into plaintiffs.  There is jockeying between groups of plaintiff lawyers, as well as attempts to align with state attorneys general and other government officials who can often bring plenty of leverage to the table. 

This all takes time to grow from the concept of a new litigation to a litigation that will produce trial verdicts and/or large settlements, even if prompted by a development independent of the plaintiffs’ bar (e.g., a published study, a recall).  By way of example, the Social Media MDL pending in the United States District Court for the Northern District of California was established by the Judicial Panel on Multidistrict Litigation in 2022, following petitions to consolidate 84 separate cases around the country.  In re: Social Media Adolescent Addiction/Personal Injury Prods. Liab. Litig., 637 F. Supp. 3d 1377 (J.P.M.L. 2022).  This means that the efforts to create the litigation probably started by 2020.  So, those in the tech industry who are only now starting to realize that their companies may be targets under new theories are facing opponents with more than a five-year head start.  While the theories they advance are often unproven, the amounts at stake encourage aggressive litigation from firms backed by investors who treat litigations as venture capitalists treat their portfolios.

Sympathy and Empathy

The legal aphorism that “bad cases make bad law” can be morphed to “sympathetic plaintiffs tend to get rulings to keep their case going.”  Judges, whether appointed for terms, appointed for life, or elected, are people too.  They can have sympathies and be concerned about how they are perceived for ending a sympathetic plaintiff’s case on technical or legalistic grounds.  Parents of children who committed suicide are incredibly sympathetic.  Although there may be good “book” arguments about contractual limits, assumption of the risk, or contributory negligence, for instance, making them under jury or public scrutiny is risky.  Trying to defend litigation on a case-by-case and factual basis is far more expensive that a defense based on legal arguments that apply across cases.  It is not hard to see why these plaintiffs’ cases, or those of state AGs, might not be dismissed even though current product liability or consumer protection laws do not support their claims or because the defendants have one or many solid legal defenses.  The trial court dynamic, particularly in the context of a bellwether trial or other early trial in serial litigation, favors tilting the table for the plaintiff.  If the defendant wins the trial after the plaintiff got the breaks, then there is no real appeal chance.  If the plaintiff wins, regardless of which side got better pretrial rulings, then there will be an appeal by the defense, and an appellate court can decide the big issues relating to expanding liability.  So, the sympathetic plaintiff leading the way in litigation can expect to get rulings that go her way to expand liability and diminish defenses.  Playing off of that expectation is part of the plan.

Without speaking directly to the defendants in the social media litigation and their alleged conduct, it is easy to see how they are relatively easy targets before a jury.  They have lots of money.  They make money without selling something tangible or, arguably, with inherent value to society, despite near ubiquity.  Many people do not understand the technology behind these businesses but tend to expect all technological challenges to be foreseeable and fixable.  In court, their executives and potential fact witnesses may communicate with technical jargon or in abstract terms that jurors find off-putting.^[3]  Their leaders are often perceived from their public appearances, statements, and press as being different, for better or worse.  They are not going to provoke more sympathy or empathy than plaintiffs who are the parents of suicide victims, for instance.  Juries have a much harder time seeing things from the perspective of a tech company or its executives than putting themselves in the shoes of the grieving parent.  Getting jurors to acknowledge the role of individual or parental responsibility when interacting with social media or other software-driven tech is a hard sell.

The Law Is Not Static

Even without the modern litigation machine driving expansion, a pattern of expanded liability over time has played out in product liability and other large-scale tort litigation many times.  For online and other software-based technology companies now in the crosshairs, it appears that what is past is prologue. U.S. tort law traces to a narrow claim related to neighboring properties in a Fifteenth Century English case commonly referred to as the Case of Thorns.  Ninety-eight years ago, future Justice Cardozo injected the concept of foreseeability as a limitation on liability for downstream consequences of negligent acts.^[4] 

Over time, the number of tort liability theories, range of compensable injuries, and types of defendants have increased dramatically, often driven by unusual cases and perceived gaps in what tort law allowed.  Taking medical monitoring as an example, the case recognized as inventing the theory was Friends for All Children v. Lockheed Aircraft Corp. a case involving Vietnamese orphans who survived a plane crash that allegedly put them at risk of neurological disorders.^[5]  A three-judge D.C. Circuit panel, which included two decidedly conservative judges (Robert Bork and Kenneth Starr), affirmed the district court decision endorsing a court-supervised medical monitoring program for an uncommon set of plaintiffs.  Over the next 15-20 years, a number of courts adopted the new claim in a range of generally expanding situations.  Along the way, the threat of nearly limitless medical monitoring damages drove several large settlements, which often included more defined monitoring programs than what the plaintiffs sought. 

While the appropriateness of the proposed monitoring varied from case to case—e.g., was it different than normal care and would it prevent worse outcomes?—the core issue was typically that the plaintiffs (typically a class) lacked a present injury apart from the exposure to the toxin, prescription drug, etc.  Eventually, the trend reversed and the long-unquestioned requirement that a tort plaintiff must have a present injury became the foundation for rulings by a number of state high courts rejecting medical monitoring without present physical injury.^[6]  A number of states still accept the claim, though, sometimes only in limited circumstances (e.g., environmental exposure in Florida).

Another example is public nuisance, which is still largely in legal flux.  Before mass opioid litigation began in 2017, cities, counties, and States had generally failed in their prior attempts to shift a range of governmental expenses to a variety of defendants and industries that sold products allegedly bad for public health.  Traditionally, public nuisance claims had to relate to the defendant’s impact on everyone’s property by how it operated a business on its property.  By contrast, claims based on damages caused allegedly by products such as guns, sodas, and paint had to be brought based on product liability principles, which meant that governments were not proper plaintiffs.  Enough appeals are still pending that the full story has not been written on the expansion of public nuisance to accommodate governments suing drug manufacturers, distributors, and dispensers for the costs of governmental services that allegedly increased due to illegal drug use.  There have been, however, enough rulings expanding the theory to force some very large settlements before some reversals on appeal.^[7]  To the extent there is a guiding principle for these and other expansions of liability, it often relates to who should bear the cost of medical expenses, remediation, etc., between the plaintiffs, the defendants, insurers, and society at large.^[8]  The knee jerk response of a judge or jury, whether in the context of social media litigation, litigation over loss attributed to software, public nuisance litigation related to broad societal harms, or litigation over traditional products, is almost always going to make the profit-making seller responsible.  While the weighing of these concerns is typically the province of legislatures, courts often get the first say.

Icebergs and Dominos

The old saying about icebergs is that the ice below the water is always much bigger and more dangerous than what you can see above the water.  Given the extent of the visible ice of social media litigation, whatever comes next will be big.  We also see the expansion in theories of liability affecting other fact patterns, alleged plaintiff injuries, and new defendants.  The dominos will continue to fall before any true correction occurs.  The impact on product liability claims against other software and software-driven “products” is obvious.  Emboldened plaintiff lawyers can easily identify other targets that can no longer rely on contractual provisions limiting liability.  Finding clients who have plausibly been injured by the software and online platforms is harder, but they can also benefit from the likely loosening of the standards for what constitutes a legally cognizable injury.  Hours spent scrolling or gaming allegedly because the design of the application/program made it “addictive” is not an injury in a traditional sense, even if other aspects of the plaintiff’s life suffered as a result.  Being subjected to hurtful comments or being made to feel bad about yourself by an AI-driven chatbot are not either.  Even in the case of health monitoring apps that are regulated by FDA as medical devices—“software as a device” as opposed to general wellness applications that are not—many of the possible failures are going to result in scares or even transient additional medical care, but those are not injuries.^[9] 

The need to spend money to prevent a physical injury has historically only been compensable (outside of the minority position on medical monitoring) if the plaintiff already has an injury that needs to be mitigated.  There will be some concrete and tangible injuries, both in a state law and standing sense, connected to these high-tech defendants, but many more claims will be based on injuries that are intangible and subjective like the intangible good on which they are based.  As a practical matter, it would be unusual for a court to expand product liability law to facilitate suing over alleged design defects in software only to block a particular plaintiff’s recovery based on a ruling that her alleged psychological trauma is not a compensable injury.

The Future Is Not Set

While we are not predicting changes as radical as those driven by the AI platform Skynet in The Terminator films, the legal framework for tort liability is changing.  These changes are likely to go beyond cases against social media companies and their close cousins.  There are some obvious parallels to possible litigation against other industries where one could allege that what they sell is designed to be addictive or that they have taken insufficient steps to protect users/purchasers from third-party criminal or bad acts.  Some of these industries, even those that sell products under traditional tort law definitions, have been relatively free of significant tort liability.  That could change.  Just as the definition of injury expands—and it should be easy to imagine claimed injuries distinct from physical injury or damage to property—the defendants in these cases may be unable to rely on the traditional defenses based on the plaintiff’s assumption of the risk or criminal act.

Consider an actual theory of liability advanced in opioid litigation against manufacturers, distributors, and pharmacies:  (1) The defendants were negligent in allowing too many addictive prescription opioid medications to end up in certain communities; (2) the availability of these medications helped facilitate the illegal use of prescription opioid medications, prescription benzodiazepines, opiate street drugs brought in by cartels, and even completely different street drugs like methamphetamine; (3) the illegal use increased local governmental costs for social services and a range of other traditional local government functions; and (4) the cost of all remediation should be borne by the defendants, even though there was no evidence linking any specific shipments or legal dispensation of the prescription opioid medications to the parade of horribles from illegal drug use.  Now, instead, start the circuitous liability theory with a popular application, AI-driven tech, or even a food item or consumer product and connect it to a multifactorial social ill.  Also consider that the historic protection from product liability for innovators and researchers who did not sell the drug or device may not apply to product liability or public nuisance lawsuits over software and AI that rope in a range of defendants hoping to find enough deep pockets.^[10] 

Those suits, seeking very high compensatory damages and/or costly injunctive relief, are coming.  The social media litigation, and to some extent the opioid litigation, made them more of a reality and threat.  How the targets of this litigation respond, in terms of the litigations themselves, legislative efforts, and adjustments to business practices will go a long way to determining whether the increased liability will inhibit technological innovation or improve it.

[1] For a discussion of caselaw through early 2023, see E. Alexander, G. Stegmaier, et al. “Predicting Risk and Examining the Intersection of Traditional Principles of Product Liability Laws with Digital Health,” in International Comparative Legal Guide’s Digital Health, 2023.

[2] See E. Alexander, “Digital Health Liability Law in Flux, ”Drug & Device Law Blog, Jan. 22, 2026.

[3] See A. Grande, US Judge Duo Urge Simplicity in Complex AI, Privacy Fights, Law360, Mar. 30, 2026.

[4] Palsgraf v. Long Island R.R., 162 N.E. 99 (N.Y. 1928) (“Negligence is not a tort unless it results in the commission of a wrong, and the commission of a wrong imports the violation of a right. . . . One who seeks redress at law does not make out a cause of action by showing without more that there has been damage to his person. If the harm was not willful, he must show that the act as to him had possibilities of danger so many and apparent as to entitle him to be protected against the doing of it though the harm was unintended.”).

[5] 746 F.2d 816 (D.C. Cir. 1984).

[6] See, e.g., Baker v. Croda, Inc., 304 A.3d 191 (Del. 2023); Berry v. City of Chicago, 181 N.E.3d 679 (Ill. 2020); Caronia v. Philip Morris USA, Inc., 5 N.E.3d 11 (N.Y. 2013); Donovan v. Philip Morris USA, Inc., 914 N.E.2d 891 (Mass. 2009); Henry v. Dow Chemical Co., 701 N.W.2d 684 (Mich. 2005); Wood v. Wyeth-Ayerst Labs, 82 S.W.3d 849, 859 (Ky. 2002); Hinton v. Monsanto Co., 813 So. 2d 827 (Ala. 2001); Temple-Inland Forest Products Corp. v. Carter, 993 S.W.2d 88 (Tex. 1999); see also Metro-North Commuter Railroad Co. v. Buckley, 521 U.S. 424 (1997) (applying the Federal Employers Liability Act); La. Civ. Code art. 2315(B) (“Damages do not include costs for future medical treatment, services, surveillance, or procedures . . . unless . . . directly related to a manifest physical or mental injury or disease.”).

[7] Harrington v. Purdue Pharma L.P., 603 U.S. 204 (June 27, 2024), details a global settlement with one of the main targets of the litigation, which had been endorsed by a bankruptcy court.  A discussion of much of the law is discussed here.  It took six years for the Ohio Supreme Court in In re National Prescription Opiate Litigation, 265 N.E.3d 1 (Ohio 2024), to rejection the acceptance of a product-based public nuisance claim by the MDL court in In re Nat’l Prescription Opiate Litig., MDL No. 2804, 2018 WL 6628898 (N.D. Ohio Dec. 19, 2018).  The Oklahoma Supreme Court took about two years to reverse a $465 million verdict by rejecting the public nuisance claim in State v. Johnson & Johnson, 2021 OK 54 (Okla. Nov. 9, 2021).  By contrast, the Fourth Circuit revived a West Virginia products-based public nuisance claim in City of Huntington v. AmerisourceBergen Drug Corp., 96 F. 4th 642 (4th Cir. 2025), which had been rejected on the law and the facts after a bench trial in City of Huntington v. AmerisourceBergen Drug Corp., 609 F. Supp. 3d 408 (S.D.W. Va. 2022). 

[8] In prescription drug product liability litigation, appellate courts created claims for “negligent marketing” and “failure to recall” as a direct reaction to the failure of the estates to prove liability under established product liability claims even with deaths tied to the use of withdrawn drugs.  See Lance v. Wyeth, 85 A.3d 434 (Pa. 2014) (Pennsylvania law); Wimbush v. Wyeth, 619 F.3d 632 (6th Cir. 2010) (Ohio law).

[9] The growth in this area, arguably outpacing regulations, can be seen in how the three-quarters of the FDA list of featured novel devices from 2025 are either software or driven by software.  When the regulatory framework is unsettled, it increases the chance that a plaintiff can enhance their claims with an allegation that an FDA-regulated “product” was sold without appropriate authorization and was thus “misbranded” under the FDCA.

[10]  Under both common law and state statutes, product liability applies to manufacturers, sellers, and distributors of products.  Efforts to expand liability to deeper pockets with no direct connection to the product that allegedly harmed the plaintiff have generally been unsuccessful.  Largely in response to federal preemption for most product liability claims against manufacturers of generic drug, plaintiffs pushed an illogical theory:  that a branded drug manufacturer should be liable for injuries alleged caused by taking the generic drug made by its competitor.  Almost two decades after this expansion of product liability was first accepted in Conte v. Wyeth (Cal. Ct. App. 2008), only Massachusetts has joined California in allowing it.  T.H. v. Novartis Pharmaceuticals Corp., 407 P.3d 18 (Cal. 2017); Rafferty v. Merck & Co., 92 N.E.3d 1205 (Mass. 2018).  Playing out something that we may see with expanding liability for social media and software, the Alabama legislature took about a year to undo the Alabama Supreme Court’s acceptance of the theory in Wyeth, Inc. v. Weeks, 159 So. 3d 649 (Ala. 2014) (amending a replacing a similar decision from 2013).  Four years later, the Alabama Supreme Court ruled that the legislature did what it intended.  Forrest Labs. v. Feheley, No. 1180387, 296 So. 3d 302 (Ala. 2019).