Meta and the states are talking settlement while the jury is still sitting

Talks are reported to be under way in Oakland, and Nevada’s attorney general has trailed a major consumer protection settlement.


Meta Platforms technology company displayed on mobil device

Meta Platforms technology company displayed on mobil device

Image Credits Credit: gguy / Shutterstock.com

Meta and state attorneys general have held settlement discussions amid the youth social media harm trial now underway in federal court in California, according to Bloomberg.

The case is the largest of its kind, with 29 states pursuing the company and four of them, California, Colorado, Kentucky, and New Jersey, party to the trial that opened this month in Oakland.

Mid-trial settlement talks are not unusual in themselves. They tend to signal that at least one side has revised its estimate of how the evidence is landing, and that this particular trial has produced several days’ worth of internal documents that Meta would presumably rather not be read aloud.

A separate signal arrived from Nevada as Attorney General Aaron Ford announced a press conference regarding what he described as a major consumer protection settlement involving a leading technology company, which he did not name.

Nevada is among the states suing Meta, and the timing invites the connection, though no party has said the two are the same matter.

Timing cuts both ways in a jury trial. Meta gains from removing the risk of a runaway verdict, while the states gain leverage from every day the evidence stays in front of a jury, which is why mid-trial talks are usually a negotiation about who blinks first.

The claims run along two tracks. The states allege violations of consumer protection law, and separately breaches of the federal Children’s Online Privacy Protection Act through collecting children’s personal data without parental consent and using it for AI training.

That second element is the one with implications beyond this case. A finding on whether children’s data gathered without consent can lawfully be used to train models would reach well past Meta, into every company that scraped consumer platforms for training material.

Jury selection began on 12 August, and a federal appeals court declined to halt the proceedings days beforehand. Meta had sought that intervention, which is worth remembering when reading its subsequent appetite for a negotiated exit.

The company has not commented on the settlement reporting. Its position throughout has been that social media addiction is not a recognised psychiatric condition and that it has worked substantially to protect children on its platforms.

Settling would also stop the flow of internal documents into the public record, which for a company facing further litigation on the same facts is worth something independent of the money. Evidence admitted in one trial has a way of turning up in the next.

The damages figures involved explain why a settlement would appeal. The four trial states are seeking roughly $200bn, and Meta has said elsewhere that states want as much as $1.4 trillion in penalties, a number that exists to be argued down rather than paid.

Any settlement would also have to satisfy a large and varied group. Twenty-nine attorneys general have their own political calculations, and a deal acceptable to one state’s electorate is not automatically acceptable to another’s, which is a familiar complication in multi-state actions.

Precedent from the tobacco and opioid multi-state actions suggests these deals tend to combine a payment with conduct commitments, and the conduct half is usually where the negotiation actually happens. What Meta agrees to change about teen accounts would matter more than the headline figure.

Meta faces parallel proceedings regardless. A separate jury in Tennessee is hearing a case on Instagram’s addictive design, and settling in Oakland would not close that front.

For now, the trial continues, with the Instagram head Adam Mosseri on the stand this week. Also, talks reported mid-proceeding frequently produce nothing, and the only reliable indicator that this set has succeeded would be the case ending before the roughly five weeks it is otherwise expected to run.

Get the TNW newsletter

Get the most important tech news in your inbox each week.