Meta agrees to $17bn child safety settlement: What changes for teens on Instagram and Facebook?

What was Meta accused of?

The settlement comes amid a much wider set of lawsuits over alleged harms caused by social-media platforms to children and teenagers. In March, a Los Angeles jury in K.G.M. v. Meta et al held Meta and Google liable for $6 million in damages after finding that features of their platforms contributed to a young user’s mental-health harms.

Central to these cases is the design of these platforms themselves. Plaintiffs have pointed to endless feeds, recommendation algorithms, notifications and social metrics such as likes and views as features designed to keep young users engaged. In the KGM case, lawyer Mark Lanier compared social-media platforms to dopamine-seeking “slot machines”.

The states’ litigation forms part of a federal Multidistrict Litigation (MDL) encompassing roughly 3,000 lawsuits, which coordinates similar federal cases for discovery, common legal questions and test trials.

The states alleged that Meta designed Facebook and Instagram to keep young people compulsively engaged, knew about serious risks and misled the public about them. They also accused Meta of collecting personal information from children under 13 without complying with the Children’s Online Privacy Protection Act (COPPA).

What has Meta agreed to change?

For the first five years of the settlement, Meta has agreed to impose a combined two-hour daily limit on Facebook and Instagram for users under 18, with direct messaging excluded from the limit. A verified parent can change the limit.

Teen users will also be blocked from most parts of the platforms between midnight and 6 am, with most push notifications disabled between 10 pm and 7 am. Notifications will also be muted between 8 am and 3 pm on school weekdays. Direct messaging will remain available during the nighttime block.

The agreement also provides for a non-personalised feed option, hides likes and reactions by default, restricts cosmetic-procedure filters, and requires stronger age-assurance, parental controls and safeguards around harmful content.

These provisions address an issue that had already surfaced during the earlier litigation: whether safety measures work when teenagers have to activate them themselves. At the KGM trial, an internal Meta document cited in court showed that only 1.1% of teen users used its existing daily-use limit.

The settlement makes some of these safeguards default settings, while others remain optional. Robbie Torney, Senior Director of AI Programs at Common Sense Media, said that distinction could determine how widely they affect teenagers.

“I’d expect the time and night-mode limits to have a real effect simply because they’re universal by default,” Torney told The Indian Express. “The recommendation and engagement changes will likely reach fewer teens because a parent must go into the settings menu to enable them.”

According to Carolina Rossini, Professor of Practice at UMass Amherst’s School of Public Policy, the settlement is the first instance in the US where a major platform has been forced to change the “architecture of its product”, rather than its policies alone.

“The dollar figure will get the headlines, but the consequential terms are those changes in design,” she told The Indian Express. “A safety tool a teenager has to find and turn on is fundamentally different from a constraint built into the product itself, and this agreement finally recognizes that distinction.”

Will these safeguards reduce harm to teenagers?

As part of the settlement, Meta will have to hire an independent auditor to assess its compliance with the safeguards for five years.

The auditor will have access to non-privileged information, personnel, systems and records relevant to assessing Meta’s implementation of the safeguards, including raw and aggregated data, internal documents and communications. Whether the measures are effective in reducing harm remains to be seen.

“Enforcement lives or dies on verification,” Rossini said, noting that the five-year audit period is shorter than the ten-year settlement timeline.

“Meaningful enforcement requires an auditor Meta doesn’t select, access to internal engagement and usage data rather than company self-reports, reproducible findings, and defined penalties for noncompliance,” she told The Indian Express.

Attorneys representing multiple states pose for a photograph outside court after after Meta reached a settlement to end the landmark teen social media addiction trial in California, Wednesday, Aug. 26, 2026. (AP)

Rossini also pointed to the possibility that teenagers could shift their activity to direct messages, which are excluded from the daily time limit, or to platforms outside the settlement.

“Compliance and harm reduction are different questions, and we should refuse to let the first substitute for the second,” Rossini said. “The real test is whether independent researchers get enough data access to measure outcomes: sleep, wellbeing, compulsive-use patterns, not feature checklists.”

Torney also cautioned against using screen time alone to judge whether the safeguards have worked.

“Time spent on screens is a proxy, not the primary outcome that matters. The real question is what happens during the time teens are still on the app,” he said.

Torney pointed particularly to recommendation systems, which Common Sense Media has previously rated as “high risk” on Instagram. “My biggest question is how turning off algorithmic recommendations changes what teens are shown and re-engage with,” he said.

What legal questions does the settlement leave unresolved?

The settlement ends the states’ trial without a judgment on the underlying claims against Meta, while thousands of other cases alleging harms to young social media users remain pending.

The agreement itself states that it does not establish a standard of care or serve as precedent in non-participating US states or international jurisdictions.

A recurring question in these cases concerns Section 230 of the Communications Decency Act, a 1996 law that generally protects internet platforms from liability for content posted by their users. As The Indian Express has previously explained, plaintiffs have sought to get around this protection by arguing that their claims concern the platforms’ own design choices, rather than content posted by third parties.

The settlement means that question will not be decided in the states’ trial. In a Stanford Law School interview before the settlement, professor Nora Engstrom said the Section 230 issue could eventually reach the US Supreme Court. While the outcome of this “bellwether trial” would not formally bind the other 2,900-plus cases in the multidistrict litigation, it could shape settlement expectations and litigation strategies.

Could the settlement affect other social media platforms too?

Rossini argues that the agreement also gives Meta a financial interest in seeing similar safeguards adopted by its competitors. Writing in The Conversation, she noted that states that participated in the trial are guaranteed about 70% of the $17 billion (about $12.7 billion) over the decade, while the remaining 30% (about $5.3 billion) will only be released if YouTube and TikTok adopt comparable teen protections and make matching payments.

The agreement defines Snap, TikTok and YouTube as “Core Industry Members”, and Rossini described the structure as a “workable template with a built-in confession”. If all three adopt comparable safeguards, the agreement provides for stricter restrictions, including a 60-minute daily limit on each platform.

Rossini noted that the settlement itself is expected to cost Meta barely 1% of its revenue over the coming decade. But if the new safeguards affect how much teenagers use its platforms, Meta has an interest in seeing its competitors adopt them too.

Torney also pointed to the consequences of teenagers moving between platforms. “Teens don’t use one app in isolation, so if one platform has strong defaults, a teen might shift their attention to another app without those same defaults,” he told The Indian Express.

An industry-wide standard, Torney said, could include measures to protect sleep, encourage time away from platforms and support real-world connections. But the measures themselves would have to account for differences between services.

“TikTok is not the same as Instagram is not the same as YouTube, so each platform will need different product-level design changes that support safe, healthy teenage development,” he said.

Rossini also pointed to a limitation of reaching such rules through a settlement. The agreement ends the trial before a fuller public record of Meta’s internal documents and testimony can emerge.

“But regulation by settlement is regulation by accident: negotiated behind closed doors, ending discovery before the public record is complete, and expiring on a contract timeline,” she told The Indian Express.

“It took state attorneys general stretching consumer-protection law to produce rules a functioning legislature should have written years ago. The settlement proves the design-accountability strategy works. It also proves we shouldn’t have to litigate our way to it, one company at a time.