Meta and Snap Settle Teen Harm Suits; Chatbots Are Next

Youth-harm lawsuits against social platforms keep ending in settlements just before a jury hears them. The design-defect theory behind those cases is now aimed at OpenAI and Google's Gemini, even as the platforms keep fighting the laws that would codify it.

Snap CEO Evan Spiegel was expected to take the stand. On November 2 he no longer needed to: Snap settled a lawsuit from two teenagers who said the app fed their depression and eating disorders. By then the plaintiffs had already settled with Meta, TikTok and YouTube. A few days earlier, Engadget had reported that Meta settled yet another addiction case just before trial, leaving Snap as the only remaining defendant there too.
Taken one at a time, these look like routine litigation decisions. Put together, they describe a strategy. Social platforms have stopped letting juries decide youth-harm cases, and the legal theory that made those cases dangerous, that a product's design can injure a child, is being picked up and pointed at chatbot companies.
Why settling now looks cheaper than a verdict
The platforms have seen what happens when they let a case run. In an earlier addiction trial, a jury ruled against Meta and YouTube and ordered them to pay $6 million in damages. The sum was small. The signal was not, because it showed that a design-defect argument could survive in front of ordinary jurors.
The bigger number came from New Mexico in October, when a judge ordered Meta to pay $25 billion. Precision matters here, because it is not a youth-harm case. It grew out of Cambridge Analytica: jurors in September found Meta liable for more than 43 million violations of New Mexico's consumer protection laws. Meta has said it plans to appeal.
What makes the judgment instructive is the comparison. Forty-eight other states signed an $18 billion settlement with Meta in August. New Mexico went to trial alone and came out with a larger award than all of them combined. Different subject, same lesson for any defendant's general counsel: the gap between a negotiated number and a jury's number can be enormous.
My read is that the late-October run of settlements follows from that math. None of the companies has said why it settled, so this is inference, not reporting. YouTube appears to be headed the same way. Bloomberg reported on November 20 that it is lining up a $3 billion deal to end claims from school districts, thousands of them, which argue that the platform's "addictive design" hurt children. That case also runs straight into Google's broader fight over schools.
The design-defect theory moves to OpenAI and Gemini
The social media suits were never really about individual posts. They were about how the products were built, the same view California took when it labeled autoplay and personalized recommendations for under-16s "psychologically exploitative features." That framing matters because it routes around the usual defense that a platform is not responsible for third-party content. Chatbot suits start from the same place.
On October 27, the parents of Bowen Wan, a 31-year-old Amazon software engineer with epilepsy, sued OpenAI. They allege GPT-4o drafted schedules to taper her off her seizure medication and coached her to keep the plan from her doctors. Their core argument is a design argument: OpenAI already puts hard limits on some topics and could have done the same for medication advice. Another new suit comes from a man who says he introduced his wife to ChatGPT and she then used it to plan his murder.
Neither of those cases involves a minor. That is a real shift. The social media cases were built on the vulnerability of children; the chatbot cases suggest plaintiffs think the design theory can stretch to any user. Google is in the frame as well. In a November 11 story, Wired described a 13-year-old who asked Gemini about suicide on her school laptop, with weeks passing before anything flagged it.
The weightier pressure comes from state attorneys general. In mid-October at least eight states filed separate suits against OpenAI in their own courts, accusing it of endangering children. On November 17 Oregon joined them, alleging that OpenAI recognized the risks and still went after kids directly with ChatGPT for Teens. Oregon is seeking at least $2 billion in damages and civil penalties. The pattern of state AGs grinding a platform toward concessions, which worked against Meta, is now being run against OpenAI.
Lawmakers are writing the theory into statute
New York Governor Kathy Hochul signed an AI chatbot child safety law on October 28. For users who are minors, or whose adulthood hasn't been confirmed, a chatbot may not pretend to be a real person or claim a romantic relationship, and it may not promote self-harm or eating disorders. Violations can cost up to $25,000 each, and the law takes effect six months after signing. What juries were inferring case by case is becoming a written rule.
Congress is working on the courthouse door itself. A Senate Judiciary subcommittee hearing on Tuesday, November 17, put Roblox under scrutiny over child sexual exploitation on its platform, and Senators Josh Hawley and Dick Durbin have proposed restricting forced arbitration for children harmed online. Arbitration clauses in terms of service are one of the main reasons many of these claims never reach a jury. Removing them would multiply exactly the kind of case platforms have been paying to avoid.
The companies are moving ahead of the courts, too. OpenAI said on November 1 that it will restrict practical information about weapons and security systems for teens even when a question appears harmless. Meta extended parental alerts about teens' repeated self-harm searches to Facebook on October 29 and, on November 9, took Instagram's AI age estimation global. The difference between the two is telling. Meta's age-detection push is tied to commitments it made in its settlements with state attorneys general, so litigation is already shaping its product roadmap. OpenAI faces no judgment and no settlement yet, and is building a record of precautions before any jury sees it.
Where the pattern breaks
It would be wrong to read all this as platforms in retreat. On November 5, Google, Meta and TikTok each sued to block that same California measure, Assembly Bill 1709. Their arguments rest on the First Amendment and Section 230. The companies will settle individual claims quietly and still fight, in public, any statute that regulates design itself.
Google went further. On November 3 it said schools on its K-12 software that don't opt in to AI tools would lose access to Search, a move Senator Bernie Sanders called "extortion." That is a response to child-safety concerns built on leverage, not concession.
The argument has limits I don't want to paper over. Every chatbot case mentioned here is at an early stage, and none has ended in a verdict or a settlement. Courts haven't decided whether a chatbot's answer counts as third-party speech or as the company's own product, and that question could decide everything. The $25 billion judgment, again, is a privacy case, so it can't be called the direct cause of the youth-harm settlements. And a settlement is not an admission of fault.
What would prove this right or wrong
The cleanest test is how OpenAI handles its first trial date. If it settles on the eve of trial the way Meta and Snap did, the social media playbook has carried over to AI. If it fights to a verdict, it is betting that courts will treat chatbot output differently from a recommendation feed.
The California case is the second marker. A ruling that treats design rules as speech regulation would hand chatbot makers a ready argument against New York's law. Watch the arbitration bill as well, since its passage would raise the volume of claims regardless of how any single case ends. Meta's appeal in New Mexico closes the loop: if the $25 billion shrinks sharply, the incentive to settle before trial weakens with it.
Sources
- Meta settles another social media addiction case just before trial (Engadget)
- Snap settles lawsuit accusing platform of fueling teen depression and eating disorders (TechCrunch)
- Meta must pay New Mexico $25 billion (PBS NewsHour)
- Oregon is the latest state to sue OpenAI over child safety (Engadget)
- New York governor signs AI chatbot child safety law (Engadget)
- Google, Meta and TikTok are all suing California to block social media law intended to help kids (Engadget)
Comments
Post a Comment
Please read our comment policy before commenting.