Inside Meta’s Landmark Child-Safety Trial: Did Instagram Ignore Under-13 Users and COPPA Rules? | #childsafety | #kids | #chldern | #parents | #schoolsafey

Instagram says users must be at least 13. Arturo Béjar says the company knew that rule was routinely broken.

The former Meta engineering director told a federal court in Oakland that “tens of thousands” of children under 13 had surfaced in his research and that their presence on Instagram was “common knowledge” inside the company. Yet Meta, he alleged, established no serious targets for finding them, measuring the problem or removing their accounts. He described Instagram’s attitude in four words: “Don’t ask, don’t tell.”

That allegation now sits at the centre of a potentially transformative trial. California, Colorado, Kentucky and New Jersey accuse Meta of designing Facebook and Instagram to keep young users hooked, harvesting children’s data and publicly playing down risks that employees allegedly understood.

Meta rejects the charges. It says it has invested heavily in safety, built protections for teenagers and will show that the states have distorted how its platforms, research and business operate. The trial will test something larger than whether Instagram has underage users. It will ask whether a technology company can avoid responsibility for children on its platform by choosing not to look too closely.

Béjar testified that Instagram treated growth and engagement as operating priorities while safety struggled for attention. Employees working on consumer products, he said, were judged largely by metrics such as user growth, visit frequency and time spent on the platform. Those measures affected performance reviews and compensation. “If you step away from the product, they are not going to make any money,” Béjar told the court. Within that system, he said, “safety was an afterthought”.

Béjar accused the company of possessing the technical muscle to detect suspicious or deceptive accounts but failing to set comparable goals for identifying children pretending to be 13 or older. He said his research uncovered tens of thousands of suspected under-13 users and that their presence was widely known internally. His evidence supports the states’ contention that Meta benefited from younger users while avoiding the information that could force it to act. These remain allegations being tested at trial. Meta will have the opportunity to cross-examine Béjar, challenge his research and present its own executives, documents and experts.

Because US child-privacy law turns heavily on what a platform knows. The Children’s Online Privacy Protection Act, known as COPPA, restricts the collection, use and disclosure of personal information belonging to children under 13. A covered company must notify parents and obtain verifiable parental consent before collecting such data. Instagram is a general-audience service rather than a product expressly directed at small children. In that situation, COPPA applies when an operator has “actual knowledge” that it is collecting personal information from someone under 13, according to the US Federal Trade Commission. The law does not generally require every website to establish the age of every visitor. A service may initially rely on the birth date entered by a user. But if it subsequently learns that a particular user is under 13, it must comply with COPPA’s parental-notice and consent rules or delete the child’s information, the FTC’s compliance guidance explains.

That makes knowledge the battleground. The states argue that Meta had ample evidence of underage use but failed to build sufficiently aggressive systems for converting suspicion into confirmed knowledge. Meta can argue that suspected underage use is not the same as legally established “actual knowledge” about individual accounts. Béjar’s phrase is powerful because it attacks that distinction. The allegation is that Instagram did not merely struggle to determine children’s ages. It lacked an incentive to discover facts that could create legal obligations.

Yes, and that is one of the internet’s oldest age-gate problems. A child can enter a false birth date and create an account that appears to belong to a teenager or adult. The FTC acknowledges that children may register for general-audience services in violation of their terms. The dispute is over what happens next. Platforms can examine behavioural patterns, account connections, birthday posts, school information, facial-age estimates, reports from other users and linked accounts. Each method raises questions about accuracy, bias and privacy. Wrongly classifying an adult as a child can lock out a legitimate user, while intensive age checks can require collecting even more sensitive information.

Béjar’s accusation is that Meta possesses one of the world’s most sophisticated account-detection systems but failed to apply comparable urgency to underage accounts. Meta’s defence is likely to emphasise the difficulty of accurately identifying age at enormous scale and the protections it has introduced for users known to be teenagers.

No. It combines two related but legally distinct arguments. The COPPA claims concern the alleged collection of personal data from children under 13 without parental consent. The broader consumer-protection claims concern children and teenagers.

The states allege that Meta deliberately built features that encourage compulsive use, then misled users and parents about the platforms’ safety. They are targeting product mechanics such as infinite scroll, autoplay, recommendation systems, notifications and visible social-reward signals. The states argue that these features exploit vulnerabilities in developing brains and convert attention into advertising revenue.

Meta says engagement is not synonymous with addiction and disputes the claim that social media can be blamed broadly for deteriorating youth mental health. Its lawyers have also argued that the platforms provide connection, entertainment, education and support to millions of young users. The court must decide whether Meta deceptively concealed material dangers and violated consumer-protection laws, not whether every teenager who uses Instagram suffers harm.

Instagram introduced “Take a Break” to remind users that they had been scrolling for a chosen period and might want to leave the app. Béjar called it “designed to fail”. His criticism centres on defaults. Users had to activate the feature themselves and could dismiss its reminders. A safety mechanism that depends on a young user interrupting an absorbing experience to seek protection, he argued, reverses the responsibility. “A safety tool has to be on by default,” Béjar testified.

That argument reaches beyond one feature. Defaults determine whether protection covers nearly everyone or only the minority motivated enough to search through settings. Meta later moved towards stronger defaults through Instagram Teen Accounts. The company says these accounts limit who can contact teenagers, restrict sensitive content and require parental permission before users under 16 loosen important protections.

According to Meta’s account of its safety measures, Teen Accounts launched on Instagram in 2024 and subsequently expanded to Facebook and Messenger. The company says it has also introduced time-limit controls, tighter messaging rules and more information about unfamiliar accounts. The states will argue that later safeguards do not erase earlier conduct. Meta can counter that its products evolved as evidence, technology and expert guidance developed.

Béjar is not an outside critic who observed Meta from a distance. He worked at Facebook from 2009 to 2015, including as an engineering director involved in efforts against cyberbullying. He returned between 2019 and 2021 as a contractor working on safety. His criticism also became personal. In written testimony submitted to the US Senate Judiciary Committee in 2023, Béjar described how his teenage daughter received unwanted sexual advances and misogynistic abuse on Instagram. He said the experience exposed a gap between the harms teenagers encountered and the incidents Meta’s formal enforcement statistics captured.

Béjar argued that Meta concentrated on how much prohibited content it removed, while failing to measure how often young users experienced harassment, sexual advances, self-harm content or other damaging interactions. Meta can challenge his conclusions, access to decision-making and interpretation of internal events. But his employment history gives the states an insider capable of connecting product design, corporate incentives and safety discussions.

Béjar testified that Zuckerberg’s priorities shaped Meta’s internal culture and that major changes struggled to advance without his backing. Reuters reported that Béjar accused the CEO of encouraging growth and engagement over child safety. That does not by itself prove that Zuckerberg personally ordered Meta to disregard underage users or cause harm. The states must connect internal knowledge and executive decisions to particular allegedly deceptive or unlawful practices.

Zuckerberg is expected to testify, along with Instagram chief Adam Mosseri. Their evidence could determine whether the jury sees Meta’s safety record as a sustained corporate effort or a series of limited responses that left the underlying engagement model untouched.

Meta denies that it intentionally designed Facebook and Instagram to addict children. Its lawyer Paul Schmidt told the court that the company accepts it has responsibilities towards teenagers and parents and has acted both independently and in partnership with families to address safety concerns. The company points to Teen Accounts, private-by-default settings, restricted recommendations, parental supervision, time-management controls, limitations on unwanted messages and protections against nudity and sextortion.

In previous congressional responses, Meta said it had invested more than $20 billion in safety and security since 2016 and had around 40,000 people working across those functions. The figures cover overall safety and security rather than child safety alone, as Meta said that work is spread across multiple teams. The defence will also question whether the states can establish a causal link between Meta’s products and a population-wide mental-health crisis. It is expected to argue that the states are turning difficult social and scientific questions into an oversized consumer-fraud case.

The distinction is crucial: the case focuses heavily on Meta’s own conduct and product design. Section 230 of the US Communications Decency Act generally protects online platforms from being treated as the publisher of content posted by users. Meta cannot simply be held responsible in this trial for every harmful photograph, message or video uploaded by someone else. The states are instead attacking choices Meta allegedly made itself: how the apps were designed, what data was collected, what the company knew and what it told users and parents.

US District Judge Yvonne Gonzalez Rogers has narrowed parts of the litigation and is expected to remind jurors that Meta is not liable merely because third parties posted harmful content. That distinction could decide the case. If the claims are characterised as an attack on publishing or recommendation decisions, Meta’s legal protections grow stronger. If they concern deceptive statements, data collection and deliberately engineered product features, those protections may not apply in the same way.

The federal litigation brings together claims from a bipartisan coalition of 29 states. California, Colorado, Kentucky and New Jersey are presenting the first combined set of state consumer-protection claims. The federal COPPA claims involve the wider group. The states fought to proceed together so they could pool evidence and present a common account of Meta’s conduct. Meta argued that differences between state laws required greater separation. Judge Gonzalez Rogers approved a hybrid arrangement, Reuters reported. The result makes this opening trial a test case whose findings could shape the remaining litigation.

Not entirely. The court has empanelled an advisory jury. Its verdict will inform Judge Gonzalez Rogers but will not legally bind her. She will make the final decision. The unusual structure reflects the mix of legal claims and remedies involved. According to Reuters, the judge said an advisory jury could help capture community standards in a dispute affecting millions of families. The trial is expected to last about six weeks.

The states are seeking financial penalties, deletion of data allegedly collected unlawfully and changes to the way Facebook and Instagram operate. Possible remedies could affect age detection, parental consent, notifications, autoplay, infinite scroll, usage reminders and other engagement features. The final scope would depend on which claims the court accepts and what changes it considers lawful and technically workable.

The financial exposure could be enormous because penalties may be calculated across large numbers of users and alleged violations. Meta has warned that the states’ theories could generate damages reaching extraordinary levels, while the states have discussed sums running into hundreds of billions of dollars. Those figures are disputed scenarios, not an established liability. Meta is already appealing another major child-safety ruling. A New Mexico court recently ordered it to place $567 million into a teen mental-health fund and adopt stronger protections, Reuters reported.

Because the case targets the machinery that keeps people scrolling. A ruling against Meta could encourage regulators and attorneys general to pursue design-based claims against other platforms. Features once treated as neutral elements of an app could face closer scrutiny if courts conclude that companies knowingly used them to exploit children’s vulnerabilities. It could also push age assurance towards the centre of platform design. Social-media companies may have to prove that they actively look for underage users instead of relying primarily on self-declared birthdays.

But stronger age checks create their own privacy problem. Establishing whether someone is a child may require identity documents, facial analysis or other sensitive data. The challenge is to keep children out without turning every adult user into a verification file.

Meta’s age rule is simple: no Instagram account before 13. The case is about everything that happened after children clicked past it. The states say Meta knew that young users were entering, understood the risks they faced and allowed engagement metrics to outrank meaningful intervention. Meta says it built extensive protections, invested billions in safety and is being blamed for harms that cannot be reduced to one company or one app. Béjar’s “don’t ask, don’t tell” allegation gives the court a sharper question. When knowledge creates responsibility, how hard must a platform try to know?