Critics warn Meta’s $18bn teen safety overhaul leaves major dangers untouched

Young people using Instagram and Facebook will face sweeping new restrictions under an $18 billion settlement Meta reached with 48 U.S. states Wednesday.

The agreement, which still requires court approval, would force Meta to introduce stronger protections for users under 18, including a two-hour daily limit and an overnight shutdown between midnight and 6 a.m.

Meta will also hide “like” counts for younger users and strengthen technology designed to identify children’s ages, prevent those under 13 from using its platforms and automatically place teenagers into age-appropriate accounts.

The $18 billion settlement will be paid over a decade and represents a major shift in how Meta manages younger users on its platforms.

But child-safety advocates warned parents not to assume the platforms will immediately become safe.

Arturo Béjar, a former Meta engineering director who testified as the first witness in the federal trial before the settlement cut proceedings short, called the agreement a “significant milestone” but said it leaves major concerns unresolved.

“The agreement has a big problem in that it allows Meta to define harm,” Béjar said.

He argued that restricting teenagers to two hours a day does not address concerns about the recommendation systems determining what they see during that time.

Béjar also urged Meta to tackle design features that he said can leave teenagers wanting to continue using the platforms even after reaching their daily limit.

A central part of the settlement targets Meta’s ability to determine how old its users actually are.

Meta already uses artificial intelligence to identify children who may have lied about their age, including under-13s attempting to create accounts and teenagers posing as adults. But critics have questioned how reliably those systems work.

Under the agreement, Meta will upgrade its age-verification technology using its own and third-party systems, with independent audits used to measure their effectiveness.

The settlement also establishes targets aimed at reducing the number of minors incorrectly identified as adults. If Meta discovers and removes an under-13 user, it will also examine the ages of people on that user’s friends list.

Federal children’s privacy law generally requires parental consent before online services knowingly collect personal information from children under 13, and Meta’s platforms set 13 as their minimum account age.

Béjar testified last week that Meta previously operated what he described as a “don’t ask, don’t tell” approach to underage users, despite possessing sophisticated technology capable of detecting suspicious accounts.

Many of the new protections will be automatically activated rather than requiring teenagers or their parents to find and enable them.

Users under 18 will face a default two-hour daily limit that cannot be disabled without parental approval. Access will also be restricted between midnight and 6 a.m.

Other protections, however, will remain optional. These include disabling video autoplay and switching from algorithmically recommended feeds to chronological ones.

Josh Golin, executive director of the child-safety nonprofit Fairplay, said the settlement still puts too much responsibility on parents.

“We are disappointed that the settlement does not turn off by default recommendation algorithms that connect kids to predators and send young people down dangerous rabbit holes,” Golin said.

The trial scrutinized the effectiveness of Meta’s existing safety features. Instagram chief Adam Mosseri was questioned about the platform’s “take a break” feature, which initially saw limited use among teenagers.

Meta later made the feature a default for its teen accounts, although users can still dismiss the notification and continue scrolling.

Researchers cautioned that announcing new safeguards and making them work effectively are two different things.

Yaël Eisenstat, director of policy and impact at the Cybersafety Research Center and a former Facebook employee, said it remains too early to know whether the settlement will produce all the changes advocates want.

But she said the agreement demonstrates that significant changes to social media platforms are technically possible.

Critics have long argued that Meta could have implemented stronger protections earlier but chose not to alter lucrative engagement-driven features.

“There’s no singular silver bullet for fixing all of this,” Eisenstat said. “I’m just glad that at least we are catching up a little bit now.”

The settlement covers 48 states, the District of Columbia and U.S. territories. Its provisions will not take effect in full unless the agreement receives judicial approval.

Meta faces tougher age checks. Two-hour limit switched on by default. Experts say the real test comes next