Meta's child-safety settlement lets the company grade its own homework

Eight days into a trial that could have produced the most damaging public record in Meta's history, the company chose to write a check instead. On August 26, Meta agreed to pay up to $18 billion and overhaul core features of Instagram and Facebook to settle claims brought by a coalition of attorneys general — led by California, joined as co-leads by Colorado, Tennessee, Kentucky and New Jersey — that the company knowingly engineered addictive products for children and lied about the harm. The trial, before U.S. District Judge Yvonne Gonzalez Rogers in the sprawling federal multidistrict litigation over adolescent social media addiction, had opened on August 18 with testimony that was going badly for the defense. It is worth being clear-eyed about what that timing means: Meta did not settle because it decided its product was safe. It settled to stop a jury from deciding otherwise.

The numbers are real, and by the standards of consumer litigation they are enormous. Just over $17 billion of the total resolves the states' 2023 lawsuit, paid out over ten years, with the remainder covering separate claims from additional states and territories that later joined the action. Illinois alone stands to receive up to $768 million. The company will also implement a raft of product changes: a default two-hour daily cap on Instagram and Facebook use for anyone under 18, a midnight-to-6 a.m. lockout, a ban on cosmetic-surgery filters for minors, and an independent auditor empowered to check compliance.

A verdict avoided, not a case resolved

None of this happened in a vacuum. Meta walked into this trial having already lost badly once. In March, a New Mexico jury found the company liable for failing to protect children from exploitation on its platforms and ordered it to pay $375 million in that state's separate consumer-protection case. When the Oakland trial opened, jurors heard from Arturo Béjar, the former Meta engineering director who has spent years documenting — internally and then publicly — how leadership treated teen-harm metrics as a growth problem rather than a safety one. Instagram head Adam Mosseri was questioned about the actual adoption rate of the company's own "Take a Break" feature, which the company had publicly promoted as evidence of good-faith reform. That is the record a settlement makes disappear. A trial produces a public judgment on the facts; a settlement produces a press release that both sides get to write.

Meta's own framing of the deal makes the point. The company's statement, delivered by chief legal officer C.J. Mahoney, describes a "new set of rules governing teens' use of social media" that will "empower parents" — language that concedes nothing about what the platforms did to make such rules necessary in the first place. That is not incidental. It is the entire function of a pre-verdict settlement: it buys the right to describe your own conduct on your own terms.

The tobacco playbook, minus the enforcement

Commentators reaching for a precedent have landed, correctly, on the 1998 tobacco Master Settlement Agreement, which required cigarette makers to fund anti-smoking campaigns and curb marketing to minors in exchange for ending state lawsuits. That comparison should worry Meta's critics more than it comforts them. States collected more than $200 billion from the tobacco deal, yet a generation later, the Campaign for Tobacco-Free Kids has documented that states have spent barely 3 percent of that money on the tobacco-prevention and cessation programs it was meant to fund; most went to plug unrelated budget holes. Nothing in the Meta agreement guarantees a different outcome for the "youth online safety initiatives" its money is earmarked for. A ten-year payment schedule to fifty-odd state treasuries, with no binding national standard for how the funds must be spent, is an invitation to repeat that history.

"You only get like two hours of alcohol or two hours of cigarettes a day, but it's still as bad for you because of what's getting delivered," said Arturo Béjar, the former Meta engineering director, of the settlement's reliance on time limits rather than changes to what the algorithm actually shows children.

Béjar's larger point is the one defenders of the deal keep sidestepping. Josh Golin, executive director of the children's advocacy group Fairplay, put it more bluntly: the settlement does not turn off, by default, the recommendation algorithms that steer young users toward compulsive engagement and, in the worst documented cases, predatory contact. A two-hour cap on how long a teenager can be on Instagram says nothing about what Instagram does with those two hours. Time limits are the easiest thing for a company to concede because they are the easiest thing to advertise; they are also the least disruptive to the underlying business, which still depends on maximizing attention per minute delivered.

Who audits the auditor

There is a second, quieter concession buried in the agreement that deserves more scrutiny than it has received. As part of the settlement, participating states agreed not to bring COPPA claims tied to Meta's use of children's data specifically for building its new age-verification systems — even though the Children's Online Privacy Protection Rule exists precisely to bar the collection of that data without parental consent. Meta is barred from using the data for ad targeting, and an independent auditor is meant to police the line. But that auditor answers to the terms of a settlement Meta helped negotiate, not to an act of Congress or an FTC rule, and the agreement is notably vague on how long the data can be retained or how cleanly it can be walled off from Meta's other systems. Attorneys who have reviewed the carve-out warn it could discourage exactly the kind of future enforcement — by the FTC, which was not a party to this deal — that would otherwise backstop it.

None of this means the settlement accomplishes nothing. Default overnight lockouts, hidden like-counts, and faster response times on abuse reports are concrete, measurable changes that will affect how millions of teenagers experience these products within months, as Attorney General Rob Bonta has promised. Common Sense Media's Jim Steyer is right that the case reflects a genuine shift in how the public — and juries — now view platforms that spent a decade insisting their products were harmless. That shift is real and it is durable, regardless of what happens with this particular settlement.

But an $18 billion payout, spread over a decade, against a company that posted $60.8 billion in revenue in a single quarter this year, is not the deterrent its size suggests. It is a cost of doing business, paid to avoid a jury verdict whose findings — on what Meta knew, when it knew it, and how it weighed teenage mental health against engagement metrics — would have lived on as public record and precedent for the state and school-district cases still coming. Settling made that evidence go away. States got a number; Meta got the narrative and a lawful path to keep the data it needed most. Congress, which has spent years failing to pass baseline federal legislation on children's online safety, should treat this settlement as the floor it is, not the solution it is being sold as, and give the FTC an enforceable national rule that does not expire, sunset, or depend on which state sued first.