A new study from researchers at New York University and Northeastern University found that at least half of the safety features meant to protect children on social media platforms fail to do what they promise. Published by Heat Initiative and Cybersafety Research Center, the study tested 86 features across Instagram, Snapchat, TikTok and YouTube. Every single platform had a failure rate of at least 50 percent for its advertised protective features.
Those features include things like blocking adults from messaging children and stopping underage accounts from seeing harmful content. As reported by Engadget, the findings raise serious questions about how much protection these platforms actually provide to younger users, especially as tech companies face growing pressure from regulators and the public to do better.
To run the tests, researchers created dummy accounts designed to mimic children of different ages, alongside adult accounts. They then looked at three different situations: a child using the platform normally, a teen actively trying to get around a safety feature, and a “malicious adult actor” trying to bypass protections on a teen’s account. A feature was counted as a failure if it was too hard to find to be useful in real life, if it didn’t do what it claimed to do, or if it was missing from the platform entirely.
Some of the specific failures uncovered are hard to ignore. On Snapchat, adult accounts were able to search for, find, and message child accounts with no restrictions at all. On TikTok, the platform was suggesting anorexia-related searches to teen accounts. These aren’t edge cases or minor bugs. They point to gaps in the core protections these platforms say they offer.
The platforms pushed back. Spokespeople for Snap, Meta and YouTube all contested the study’s findings in statements to the New York Times, which also said it was able to replicate the results independently. A Meta spokesperson told Engadget that thanks to Instagram’s Teen Accounts, teens are “seeing less sensitive content, experiencing less unwanted contact, and spending less time on Instagram at night.” Meta also argued the study’s authors “include vague claims that our features are broken but, in the vast majority of cases, either misrepresent those features or fail to provide any examples or evidence.”
This study lands at a particularly difficult moment for social media companies. Platforms are facing lawsuits from school districts and individuals who say their products caused direct harm to young users. Multiple countries are also moving toward outright bans on social media for children. Australia recently doubled the maximum financial penalty for companies found breaking its child social media ban, a sign that governments are losing patience with voluntary compliance.
The broader pattern here is one of a widening gap between what platforms say their safety tools do and what independent researchers find when they actually test them. That gap matters because parents, schools and policymakers often make decisions based on the assumption that these features work. If half of them don’t, the entire framework of platform self-regulation looks a lot shakier than the companies would like to admit.
Correction, June 29, 2026: An earlier reference to an Instagram feature that implied it was failing to prevent contact between teens and adults has been removed. The study itself acknowledged the feature was working as intended. Engadget regrets the error.