Meta Settled. The Social Media Mass Tort Did Not.

Meta Settled. The Social Media Mass Tort Did Not.

For eight days in Oakland, the legal world finally got the Meta trial it had been waiting years to see. The trial started on August 18 before District Judge Yvonne Gonzalez Rogers in the Northern District of California. A coalition of state attorneys general alleged that Meta designed Facebook and Instagram to almost unwillingly keep children engaged, misled the public about the risks of engagement, and collected data from children without parental consent. The case was part of MDL No. 3047, the social-media addiction litigation that, as of early August, had more than 3,100 actions pending. The trial was supposed to run six weeks. It lasted eight days. On August 26, Meta agreed to a settlement worth up to $18 billion. The trial stopped immediately.

This was a big deal. New Jersey Attorney General Jennifer Davenport described it as the largest state consumer-protection settlement outside the Big Tobacco cases of the 1990s. California Attorney General Rob Bonta called it a “milestone moment” and said the states had obtained the changes they wanted without spending years litigating the matter. Colorado Attorney General Phil Weiser warned TikTok, YouTube and Snap that they were now “on notice.” Under the terms of the agreement, Meta admits no wrongdoing, but it agreed to pay a guaranteed $12.7 billion over the next decade, with another $5 billion tied to competitors adopting similar safety measures and resolving related claims.

The platform-based changes may matter more than the check. Users under 18 will face a default two-hour daily limit that only a parent can lift; if rival platforms agree to similar restrictions, that limit can fall to one hour. Meta will block teen use from midnight to 6 a.m., with the possibility of expanding that window to 10 p.m. through 7 a.m. if other platforms follow suit. Notifications will be shut off during the school day and overnight. Like and reaction counts will be hidden from minors, cosmetic-procedure filters will be prohibited, teenagers will be able to choose a non-personalized feed. It also must respond to at least 90% of teen reports of potentially harmful content within six hours. This is not a warning-label settlement. This deal will shift the landscape of social media platforms, and at least to this lawyer, it’s about time. 

This settlement matters because Meta is giving much more than cash, they are giving the States what they have been seeking for years. The theory has never been about one specific post causing one specific child’s injury. The States have focused on the architecture around the post: recommendation systems, infinite feeds, notifications, public popularity metrics, autoplay, age controls and the constant effort to reduce the number of moments when a user might decide to stop. Meta/Facebook have consistently argued that it takes youth safety seriously, disputes that “social-media addiction” is a recognized psychiatric diagnosis and rejects the idea that its products have caused the youth mental-health crisis. A settlement does not change any of that into an admission. But after August 26, it becomes harder to dismiss proposed product changes as fanciful ideas dreamed up by plaintiffs’ experts. Meta has now agreed that many of them can actually be implemented.

The trial ended only after the states had started putting human beings behind years of internal emails and product documents. Former Meta safety engineer Arturo Béjar, who worked at Facebook from 2009 to 2021, was the states’ first major witness. He testified that Meta’s top-down culture meant meaningful product changes generally required Mark Zuckerberg’s backing and that he had not made child safety a sufficient priority. Béjar has been a persistent critic of the company, including in testimony before Congress. His testimony was clear and helpful: an insider describing how complex decisions of child safety were made, behind the scenes. Then came Adam Mosseri, the head of Instagram since 2018. On August 25, Mosseri testified about Instagram’s “Take a Break” feature, which encouraged users to step away after spending a pre-determined amount of time. He acknowledged that very few teenagers activated it (internal data put teen adoption at around 1.8%). Mosseri defended the company’s safety work and rejected the suggestion that Instagram concealed research about youth harms.

The August 26 deal resolves the state cases; it does not resolve MDL 3047. It does nothing for the private plaintiffs. And those private plaintiffs are still very much here. Individual plaintiffs are seeking compensation for resulting depression, self-harm, eating disorders, death, and more. As of August 3, the JPML reported 3,137 actions pending in the federal MDL, and thousands more cases are coordinated separately in California state court. Beyond those claims, more than 1,200 school districts have also brought claims alleging that social-media companies helped create a student mental-health crisis that schools have been forced to pay for.

The school cases already provide an interesting preview of where this may go. Breathitt County School District, a small district in Kentucky with roughly 1,600 students, was supposed to be the first federal school-district bellwether in June. It never made it to trial. Meta paid $9 million, Snap and TikTok each paid $8 million, and YouTube paid just over $2 million, producing a combined settlement of roughly $27 million. Breathitt had sought more than $60 million, including funding for a 15-year mental-health program. Still, the first school-district test case disappeared before a jury could hear it, just as the states’ Meta trial did two months later.

TikTok has been making similar calculations. On August 3, Reuters reported that the company had reached confidential settlements in three California cases brought by teenagers from Illinois, New Jersey and Mississippi alleging addiction, depression, self-harm and eating disorders. Those cases came from the roughly 3,300 social-media lawsuits coordinated in California state court before Judge Carolyn Kuhl. TikTok did not admit liability, but when carefully selected social-media cases get close to a jury, settlements come quickly.

That does not mean a global private settlement is around the corner. The individual claims have a problem the states’ cases did not have to solve in the same way: specific causation. Family history, bullying, school, relationships, preexisting mental-health issues, COVID-era isolation and use of TikTok, Snapchat and YouTube may all matter.

That is where MDL 3047 starts to look more like a traditional mass tort. Uber’s sexual-assault litigation has already (and dramatically) shown how case value can be affected by the facts: one 2026 bellwether produced an $8.5 million verdict, another produced $5,000. Social media may produce the same spread. The valuable claims will likely require meaningful platform-use evidence, medical records, a sensible chronology and some methodology to separate the platform’s role from the rest of the claimant’s life.

For plaintiff firms, that means the next phase is probably less glamorous and more important. Intake needs to get better. Which platform? What age did use begin? How much daily use? Which features mattered? When did symptoms begin? What treatment followed? Can the claimant obtain account data? The $17 billion headline will inevitably attract firms to build up portfolios of cases using ads and referring firms. Some of these efforts will produce strong cases, but they will also produce people who vaguely remember using Instagram in middle school, so careful intakes will be important.

The settlement also puts TikTok, Snap and YouTube in a strange position. Part of Meta’s payment increases if competitors accept similar restrictions, meaning Meta has effectively negotiated a financial incentive for the rest of the industry to adopt the same rules. A two-hour Instagram limit is considerably easier to swallow if TikTok has one too. Meta therefore walks away from Oakland having paid an extraordinary amount of money but also having helped create a potential industry standard against which its competitors may now be measured. Weiser made that threat explicit: if further trials are necessary, the states are prepared to bring them.

Section 230 remains unresolved as well. The statute was designed to protect platforms from being treated as the publisher of content created by users. MDL 3047 is testing the boundary between third-party content and a platform’s own design. On August 10, the Ninth Circuit rejected Meta’s effort to take an immediate appeal and stop the trial based on Section 230. This does not mean that Meta ultimately loses the defense; but it allows the cases to keep moving. By settling the states’ trial, Meta also prevented Judge Rogers from issuing a merits ruling many lawyers wanted on how Section 230 applies to social media platforms. That issue now remains available for the private plaintiffs to keep litigating.

Meta’s settlement does not give us a clear answer as to what specific evidence will be required to establish liability. But – it does tell us that this litigation is no longer speculative. The states got billions of dollars and product changes. Breathitt County got $27 million from four platforms before its bellwether. TikTok has settled individual cases before trial. Private plaintiffs and more than a thousand school districts remain in the system. And Meta paid an extraordinary amount of money after eight days of what was supposed to be a six-week trial.  Certainly, the private plaintiffs will take longer than 8 days, but just how long, no one knows.

For the mass-tort world, that is where the story sits on August 31. The headline case settled, but the harder work has yet to begin. Plaintiffs still have to prove individual causation. The remaining platforms have to decide whether Meta just bought long-term peace or wrote the first draft of the industry’s safety standard. Either way, MDL 3047 did not end on August 26. In fact it may have just become a lot more expensive and considerably harder to dismiss.  These cases have value but be careful – there will be a lot of bad intake before finding the good.

________________________________________________________________________________________________________________________________________________________________________________________