Addictive software litigation: Early signals for casualty risk

Authors: Adam Grossman, Director - Product Management, Moody's; Taro Ramberg, Director - Marketing, Moody's

Two jury verdicts against Meta and Google mark an important development in the evolution of liability risk. In these separate cases, juries found liability on theories alleging that certain harms were associated with software platforms designed to drive prolonged engagement.

The verdicts themselves are only an early data point. The broader significance for insurers is what these outcomes suggest about where design-centered liability theories may go next, especially when AI-related technology is at their core, and how to manage emerging risk as technology and legal arguments both evolve.

 

1.      State of New Mexico v. Meta Platforms, Inc

On March 24, 2026, in the case State of New Mexico v. Meta Platforms, Inc, a jury found Meta liable for violations of the state’s Unfair Practices Act, which had been updated in 2025, based on allegations that Meta had engaged in ‘willful deception’ and ‘unconscionable business practices.’

New Mexico’s Attorney General Raúl Torrez alleged that Meta, the owner of social media platforms Facebook and Instagram, made false or misleading statements about safety for adult and child users. At the same time, it was further alleged that internal decisions to maximize engagement contributed to a range of harms, including increased exposure to harmful interactions involving minors.

After a six-week trial, the jury returned a verdict for the state, finding Meta liable on the state’s consumer protection claims. The statutory maximum penalty of $5,000 for each count was applied to 37,500 violations, resulting in a total of $375 million in civil penalties.

In a statement, AG Torrez remarked, “The substantial damages the jury ordered Meta to pay should send a clear message to big tech executives that no company is beyond the reach of the law. Policymakers and law enforcement officials across the country can help make this verdict a turning point in the fight for children’s safety. This is a watershed moment for every parent concerned about what could happen to their kids when they go online – and this victory belongs to them.”

In response, Meta has officially stated that they “respectfully disagree” with the decision and have indicated an intent to appeal.

 

2.      K.G.M. (Kaley) v. Meta, Google, et al.

A day later, on March 25, 2026, a jury at the Los Angeles Superior Court (the California Superior Court with jurisdiction over Los Angeles County), in the case of K.G.M. (Kaley) v. Meta, Google, et al., found both Meta and Google (via its YouTube subsidiary) liable on negligence-based theories a now 20-year-old plaintiff alleged were attributable to the use of their platforms.

The plaintiff described using YouTube since the age of 6 and Instagram from 9 years old and argued that specific design features, including infinite scrolling and AI-based, engagement-driven content delivery, were associated with compulsive use and related mental-health injuries, including eating disorders, depression, and suicidal ideation.

The jury also heard the plaintiff’s expert testimony citing scientific literature suggesting that certain engagement-driven design features can be associated with compulsive use and downstream harm. The jury reached a 10-2 decision, awarding the plaintiff $3 million in compensatory damages and an additional $3 million in punitive damages, allocating 70 percent of responsibility to Meta, 30 percent to Google. TikTok and Snap were defendants earlier in the case but settled before trial for undisclosed amounts.

 

Implications for (re)insurers

What is critical for (re)insurers to understand is not the size of these verdicts, but the underlying theory of harm, which in these cases involved claims tied to allegedly engagement-driven software design.

Core allegations in these matters, and in the more than 4,000 similar current cases, generally do not center on ‘social media’ as a category; instead, plaintiffs generally allege that certain design features were selected to increase engagement and may draw on behavioral mechanisms discussed in the scientific and policy literature.

In many of these cases, plaintiffs argue that algorithmic optimization plays a central role in how engagement is created and sustained, which is why the potential relevance extends beyond any single platform. Analogous engagement-driven design practices appear in various forms across a wide range of industries outside social media, including video games, sports betting, chatbots, online retail, streaming services, and other consumer-facing digital products.

That industrial diversity is already visible in the ongoing litigation. According to Moody’s Litigation Tracker, part of the CoMeta™ casualty catastrophe modeling platform, as of March 26, 2026,  1,168 cases are pending in coordinated proceedings in the Los Angeles Superior Court jurisdiction targeting social media companies. A separate set of cases for the Court targeting video game companies includes 40 lawsuits pending against seven different companies.

More broadly, addictive software design actions around the country now include more than 4,000 cases targeting 166 different companies across a wide range of industries, including online gaming, sports betting, and chatbots.

Now that these two initial verdicts have made headlines and the current scope of the litigation is clearer, the next question for (re)insurers goes beyond these two verdicts and to how the underlying theories translate into coverage, accumulation, and portfolio exposure. This shifts the focus to the underlying basis for liability and what it could mean for portfolios, both for these cases and for others already in the pipeline or still to come.

The New Mexico case is more straightforward from an insurance-response perspective. The $375 million award consists of civil penalties under state law, which often function more like regulatory fines than covered damages under standard liability policies.

To the extent that the joint federal lawsuit filed by 33 states in the Northern District of California, and separate lawsuits from another 9 states and D.C., all proceed under similar legal theories, the direct insurance impact may be limited, though outcomes can vary by jurisdiction and policy language.

In an early coverage decision on February 27, 2026, the Delaware Superior Court, in the case of Hartford Casualty Insurance Co. v. Instagram, LLC, Judge Sheldon K. Rennie ruled on partial summary judgment that certain insurers had no duty to defend Meta in the thousands of ‘social media addiction’ and child safety lawsuits across the country.

The judge ruled that the allegations underlying the major litigations in California concerned allegedly deliberate conduct, and that the alleged harms did not arise from an ‘accident’ and therefore fell outside the scope of coverage under the policies at issue, including the duty to defend. Meta has indicated it intends to appeal the ruling.

If that ruling is upheld on appeal, it could reduce defense-cost exposure for insurers across a large portion of the consolidated matters pending in California, depending on policy language and how other courts address similar issues.

 

Eight years of tracking addictive software design

For many observers, these verdicts may feel sudden, but the underlying risk appears to have been developing for years. In 2024, Moody’s acquired Praedicat, a business formed in 2012 that applies science and artificial intelligence to build a probabilistic liability catastrophe model and risk products for some of the world’s largest insurers and corporations.

CoMeta, launched in 2014, is a casualty catastrophe modeling platform that mines data from published science to identify and model casualty risks. CoMeta first identified addictive software design as a potential risk source in 2018 and has been publishing research and providing analytics to inform the market since then.

At that point, the idea that engagement-driven design could cause addiction was not widely accepted and was therefore unlikely to translate into large-scale litigation or liability. Since then, scientific evidence has matured, and legal theories have evolved, making these claims more likely to be asserted in large-scale litigation than they would have been in 2018.

In a 2022 Carrier Management article, our David Loughran estimated using CoMeta that, given current science and assuming plaintiffs could overcome various legal defenses available to the defendants, a mass litigation event centered on addictive software design would have an expected economy‑wide loss (indemnity plus defense) of approximately $15 billion, with a 5 percent probability of exceeding $70 billion.

This was not a forecast for any one case. It was a way of sizing tail risk when technological change, scientific development, broad exposure, regulatory attention, and plaintiff recruitment converge across a widening field of defendants.

For insurers, the deeper challenge is that historical loss experience alone cannot capture risks of this kind before they begin to surface in litigation. Emerging casualty risks are often byproducts of technological change, with exposure accumulating ahead of claims and legal validation. Identifying, monitoring, and managing those risks requires forward-looking analytical approaches that reflect how new technologies could translate into liability through scientific discovery and legal innovation.

In situations like addictive software litigation, the risk is beginning to emerge, but forward-looking modeling, such as CoMeta – now in its twelfth year – remains critical for understanding how exposure may sit within a (re)insurer’s book and where aggregation could be building.

As claims emerge and resolve, standard actuarial approaches that blend early loss experience with projections of ultimate loss from forward-looking models remain essential for managing reserves and cash flows. Forward-looking analytics complement those fundamentals by helping insurers interpret limited early signals in the context of evolving science, legal theories, and product design practices.

More broadly, a forward-looking approach to emerging risk management helps (re)insurers set and maintain disciplined appetites for accumulation across hazards that could, over time, translate into loss. Not every emerging risk matures into litigation or insured loss.

But managing these risks systematically, before litigation outcomes such as this week’s two widely reported cases can be added to historical data, supports more consistent underwriting and portfolio decisions and helps insurers grow sustainably through periods of rapid change.

Discover more about CoMeta, emerging risk intelligence to track over 300 emerging risks mapped to companies, industries, and policy portfolios, allowing for earlier underwriting adjustments, exposure monitoring, and reserving actions, here.