Meta’s Teen Accounts global rollout comes with a stinging rebuke from independent researchers, who claim that the company’s safety tools do not work as they are advertised. Hours after Meta disclosed that it was turning on default protections for teen users across Instagram, Facebook and Messenger worldwide, a fresh analysis found many of the safeguards held up as central to the experience could still be easily circumvented or were not effective.
Global Rollout Encounters Tough New Scrutiny
This updated foundation of Teen Accounts is meant to immediately reduce privacy and safety threats. Meta says it restricts who can contact teens, limits how easily accounts are discoverable, filters sensitive content and turns off the ability to go live for users under 16. The parental supervision features are framed as augmenting those automatic protections — which Meta describes as an industry-leading baseline.
But a report put out the same day, by an organization called Cybersecurity for Democracy, which was co-founded by a former Meta executive and whistleblower Arturo Béjar along with researchers at New York University and Northeastern University, claims that those guardrails “abjectly” fail to deliver. The report, “Teen Accounts, Broken Promises,” was released with U.S. and United Kingdom child advocacy partners including Fairplay, the Molly Rose Foundation and ParentsSOS.
What Teen Accounts Promise for Privacy and Safety
Teen Accounts broadly seek to reduce the likelihood of toxic interactions in three ways: reducing contact with unknown adults, curtailing recommendations for mature or risky content and introducing friction on features that are known to increase risk — like live streaming and late-night use. Meta has also introduced updates around its AI features for teens, limiting interactions that may veer into romantic or sexual realms.
These product decisions illustrate lessons regulators and safety groups have been pushing for years: safety-by-default for minors, fewer algorithmic roads to sensitive content and more control by parents and guardians. It’s the right direction on paper. Whether the system holds at scale, as the new testing makes clear, is an open question.
Inside the independent testing behind the new report
The researchers reviewed 47 of the 53 user-visible safety features that Meta announces for teenagers. Their report card is unkind: Thirty features — 64 percent of them — were rated “red,” for discontinued or ineffective in real-world use. Nine received “yellow” ratings for partial or inconsistent protection. Only eight were assigned a “green” rating for doing so effectively.
Examples from the testing bear out these core claims. In some cases, adult accounts could still find teen accounts — even though the intention was to restrict unsolicited contact. Teenagers could start or maintain conversations with adults who weren’t following them. Words or slurs used as insults in direct messages were recorded passing through the filters. And sex, violence and self-hatred content continued to be pushed by recommendation systems to teen profiles, the report also said.
Reporting mechanisms were also criticized. Key flows researchers said required teenagers to either receive distressing messages or be exposed to concerning content before a clear report option would surface, an approach critics say bakes harm into the system. Béjar has likened this situation to pushing a car without working brakes: By the time a teen discovers that something is wrong, it may be too late.
Meta’s response & the lack of evidence, explained
Meta disputed the findings, arguing that they mischaracterize how its tools work and how families use them in practice. The company says that teens who are put into Teen Accounts see less sensitive content, have fewer unwanted contacts and spend less time on Instagram at night. It stresses that parents are able to restrict use and supervise interactions with simple controls, and says it will continue to iterate on safety safeguards.
This difference illustrates an ongoing measurement discrepancy. Company telemetry is capable of reflecting collective progress, but auditors are paying attention to edge cases, bypass paths and consistency in enforcement — the very places safety systems tend to splinter. Until Meta opens up to reproducible third-party verification or releases detailed efficacy data by scenario, dueling claims are likely to keep feeding public skepticism.
Policy and enforcement pressures mount on platforms
The report comes at a time of increasing regulatory momentum. Some writers say that the Kids Online Safety Act should be passed and for the Federal Trade Commission, as well as state attorneys general, to use the Children’s Online Privacy Protection Act and Section 5 of the F.T.C. Act to push platforms on teen protections. For advocates in the U.K., the Online Safety Act is seen as a lever that should be made stronger when it comes to enforcement.
Advocacy groups including Common Sense Media have lambasted splashy safety launches that fall short in enforcement. Watchdogs have previously documented that teens were exposed to sexual content even when they set their accounts to more restrictive modes, and Meta has said it removed over 600,000 accounts for predatory behavior. A whistleblower from Meta called for independent audits to verify the company’s safety assertions during recent testimony in front of the Senate.
The path forward for teen safety and meaningful audits
Two challenges stand in the way of Teen Accounts to live up to their promise: age verification and regular enforcement. Default protections are leaky, if platforms can’t properly identify kids. And bad actors will find the seams if detection and reporting systems break down at the margins. Independent testing, public metrics and regular third-party audits — coupled with swift fixes — would do more than any sales pitch to earn trust.
Meta’s world-conquering ambitions shine an even brighter light on those fundamentals. The company deserves credit for making that shift to safety-by-default for teens, but the metric is results, not intentions. If the findings are borne out, then Teen Accounts won’t merely need tweaking but also substantive, testable guardrails that work before harm happens rather than after.
