A Bellwether Trial With National Stakes
On August 18, 2026, a federal trial opened in Oakland that could reshape how platforms are held liable for how they are built, not just what users post on them. Four states — California, Colorado, Kentucky, and New Jersey — are the first of 29 to try claims against Meta in the consolidated federal litigation In re: Social Media Adolescent Addiction/Personal Injury Products Liability Litigation (MDL 3047), before U.S. District Judge Yvonne Gonzalez Rogers in the Northern District of California. An eight-person jury will hear evidence, but its verdict is advisory only; the judge decides the case. The trial is expected to run six to eight weeks, with a ruling likely around October 2026.
The states' first witness set the tone. Arturo Béjar, Meta's former engineering director for protect-and-care and a long-time whistleblower, testified that internal teams relabeled findings of "addiction" as "problematic use" to produce softer statistics, and that he raised evidence of minors encountering predatory and violent content with Instagram and Facebook executives on multiple occasions without seeing meaningful product changes.
What the States Are Actually Alleging
The case, filed October 24, 2023 by a bipartisan coalition of state attorneys general, centers on two claims. First, that Meta violated the Children's Online Privacy Protection Act (COPPA), the 1998 federal law requiring verifiable parental consent before collecting personal data from children under 13. California Attorney General Rob Bonta's office says an unredacted version of the complaint shows Meta's internal data put roughly 30% of 10-to-12-year-olds on the platforms at one point — a figure hard to square with Mark Zuckerberg's 2021 congressional testimony that Meta "kicks off" underage users. Second, the states allege Meta built "psychologically manipulative" features — infinite scroll, push notifications, like-counters, and appearance-altering filters — specifically to maximize time spent, then misrepresented the resulting harm publicly. The unredacted filing also states Zuckerberg personally rejected a proposed ban on filters simulating cosmetic surgery over internal objections about mental-health effects on teenage girls.
The legal architecture here is deliberate. Rather than argue Meta is liable for what users posted, the states frame their claims around design choices and corporate statements — a route built specifically to avoid Section 230 of the Communications Decency Act, which shields platforms from liability for third-party content but does not obviously cover a company's own product-design decisions or its public safety claims. It is the same strategy that has already produced results elsewhere in this sprawling litigation: a Los Angeles Superior Court jury awarded $6 million against Meta and Google in K.G.M. v. Meta in March 2026, and a New Mexico court has now ordered roughly $942 million in combined penalties and abatement funding against Meta across two 2026 rulings, finding its platforms constituted a public nuisance.
The Case for the States — Fairly Stated
The strongest version of the states' argument does not require assuming bad intent about every product feature; it rests on the documentary record. If Meta's own researchers were producing findings about harm to minors, executives were briefed repeatedly, and public-facing statements and congressional testimony did not reflect what internal data showed, that is a garden-variety deception and negligence case, not a novel assault on how the internet works. COPPA itself is narrow and uncontroversial: it does not regulate speech or content moderation, only whether a company obtains parental consent before harvesting a child's data. Enforcing a 28-year-old statute against a company that, by the states' account, knew it had millions of underage users is not regulatory overreach — it is enforcement of a law already on the books.
Where the Theory Strains
But the scale and mechanism of relief here deserve scrutiny even from that starting point. Meta has told the court its exposure could reach roughly $1.4 trillion, against a state estimate closer to $200 billion — a spread wide enough to suggest neither side has a stable damages methodology, and closer to the sort of number Congress, not a single district judge advised by a non-binding jury, should be setting policy around. The parallel verdicts already emerging from this same MDL — $6 million in Los Angeles, $942 million in New Mexico — show state-by-state litigation producing wildly inconsistent outcomes for materially similar conduct, which is precisely the due-process and fair-notice problem that comprehensive federal legislation is supposed to solve and litigation is poorly suited to. Design features like notifications and recommendation algorithms are not inherently manipulative; they are also how a news app resurfaces breaking stories or a fitness app keeps someone on a workout streak. A damages verdict built on these features risks setting a de facto national product-design standard for one company, through one judge's remedies order, rather than through a Kids Online Safety Act or an FTC rulemaking that Congress has spent years failing to pass. If the underlying conduct is a COPPA violation, COPPA already has a statutory enforcement mechanism through the FTC and state AGs; the trillion-dollar tort theory is being used to accomplish what legislative reform of algorithmic design has not.
The more durable fix is still the one Washington keeps deferring: a federal statute that sets clear, prospective design and data rules for minors, enforced consistently, rather than forty-plus separate juries and judges retroactively pricing the same conduct forty-plus different ways.