Table of Contents
You flipped on the parental controls. You set the account to private, turned on the screen time reminders, and trusted the “teen safety” settings the app promised. New research from Northeastern University suggests a lot of that effort does far less than parents assume. When researchers set out to test 86 child safety features across Instagram, TikTok, Snapchat, and YouTube, only 35 held up. The other 51 failed, and some backfired in ways that would alarm any parent.
In one test, a research account registered as a teen searched for help, and instead of blocking harmful material, the platform started recommending it. That single finding captures the gap between what social media companies advertise and what actually protects a child scrolling alone in their room.
What the Northeastern Study Found
The study, led by Laura Edelson, an assistant professor of computer sciences at Northeastern, took a direct approach. Researchers built fake teen and adult accounts on the four platforms most popular with young people, then tested each advertised safety feature against two simple questions: does it work the way the company claims, and does a child ever actually see it while scrolling?
To count as successful, a feature had to pass both. It also had to be on by default or easy to turn on, hold up under normal teenage use, and demonstrably do its job. Out of 86 features, just 35 cleared that bar. The remaining 51 fell short, and the research sorted the failures into telling groups.
Nine features were “missing,” meaning they could not be triggered at all, even after the team followed the platform’s own instructions. Edelson pointed to an Instagram feature meant to nudge a user to reconsider before posting something cruel. When a teen account typed clear bullying at another teen, the nudge never appeared.
Thirty four features were “broken.” The tool existed but either failed at its core purpose or could be dodged with almost no effort. Many relied on blocking a list of banned words, and simply misspelling one of those words slipped right past the filter. Of those broken tools, 12 were both broken and “buried,” nonfunctional and also hidden under so many menus that an average child would never reach them. Another eight features worked but were buried so deep they were effectively out of reach.
The most disturbing result came on TikTok. TikTok states it blocks minors from searching for content about disordered eating and self harm. Edelson’s team found the opposite in practice. “It recommended that our teen account look up ‘anna food tips’ and ‘how to pretend to eat your food,’ both drawn from pro anorexia communities, alongside ‘mentally suffering,’ ‘losing yourself to mental health,’ and, most disturbing of all, ‘razor blade skin,'” the researchers reported. “These were the product’s own recommendations, served to a child, not phrases we went looking for.”
Every safeguard meant to regulate how users treat one another, the tools aimed at cyberbullying, failed on all four platforms. Only one in three tools designed to curb compulsive use, like screen time limits and break reminders, actually worked.
What the Companies and Experts Say
Northeastern reached out to all four platforms. YouTube and TikTok responded.
A YouTube spokesperson said the company has “spent over a decade building industry leading parental controls” and cited a figure that 84 percent of parents who used its supervised account tools felt more confident their child was in a safer environment. A TikTok spokesperson said teen accounts “come with over 50 preset safety features and settings automatically turned on,” added parental options through Family Pairing, and said the company’s internal review “confirms these features are working as intended.”
The research points to a different reality, and Edelson drew a clear lesson from it. “The number one lesson that I take away from this is that making the default experience the safest one, basically making safety the default, is something that works very well,” she said. The features that did pass tended to be the ones switched on automatically. Instagram sets new teen accounts to private on creation. TikTok for Younger Users, for children under 13, is view only, with no commenting or messaging. Those defaults worked precisely as a child did not have to find and activate them.
The finding lands in the middle of a broader push on child online safety. Days before the study came out, the U.S. House passed the Kids Internet and Digital Safety Act, a package that would require new safety features, restrict targeted advertising to minors, and set rules for AI chatbots. Research like Edelson’s raises a pointed question about whether features mandated on paper will function any better than the ones already advertised.
What This Means for Parents
The practical takeaway is not to panic or to hand your child an unmonitored phone in defeat. It is to stop treating a platform’s safety settings as a finished job. Here is where to put your energy.
Trust defaults more than toggles. The safest features are the ones a company turns on for everyone, like automatic private accounts and view only modes for the youngest users. Choose platforms and account types that start safe rather than counting on a long list of switches you have to find and flip.
Assume filters leak. Keyword blocks miss misspellings and slang, and recommendation engines can push exactly the content a filter claims to hide. A content filter is a speed bump, not a wall. Stay involved rather than assuming the software has it covered.
Delay the riskiest platforms for younger kids. Edelson noted that for the youngest users, removing a risky service works better than trying to fence it in. A view only account or no account at all beats a feature heavy account that fails quietly. Waiting on open, algorithm driven feeds is a reasonable call for a grade schooler.
Keep talking, and keep it open. The tools that failed most completely were the ones meant to catch bullying and harmful searches. That work falls back to the relationship. Kids who know they can tell a parent about something upsetting online, without losing the device as punishment, are far better protected than kids relying on a filter. Ask what they are seeing. Watch a few videos together. Make the conversation routine, not a reaction to a crisis.
Check the settings yourself, then check again. Some real protections exist but sit buried under menus. Spend twenty minutes in each app’s settings, turn on every restriction you can find, and revisit after updates, which can reset preferences without warning.
How the Four Platforms Compared
The study looked at Instagram, TikTok, Snapchat, and YouTube, chosen for their popularity with young people. While the research did not crown a winner, the pattern of what passed and what failed is useful for parents deciding which apps to allow and how.
The clearest successes were structural rather than optional. Instagram automatically makes new teen accounts private, so a stranger cannot follow or message a young user by default. TikTok’s version for users under 13 is view only, stripping out comments and direct messages entirely. Both worked precisely as a young child does not have to know they exist to be protected by them.
The failures clustered around anything a child had to find, activate, or rely on to catch bad behavior in the moment. Bullying safeguards failed across all four platforms. Search filters leaned on blocked word lists that a misspelling defeated. Screen time tools, the ones meant to interrupt endless scrolling, worked only about a third of the time. And on TikTok, the recommendation engine actively surfaced the exact category of content the platform says it blocks for minors.
The lesson for a parent is not that one app is safe and another is not. It is that a private, restricted, or view only account beats a standard account loaded with optional switches, whichever brand it carries.
Younger Kids and Teens Need Different Approaches
The right move depends heavily on your child’s age. For younger children, the research supports a blunt approach: remove the risky service rather than trying to build a fence around it. A view only app or no social account at all sidesteps the whole problem of features that fail quietly. Edelson made this point directly, noting that for the youngest users, taking the risky service off the table works better than guardrails.
For teens, an outright ban often backfires and pushes use underground. The stronger play is a private account on a platform with safe defaults, paired with ongoing, judgment free conversation. Teens who feel they can report a disturbing message or a harmful video without losing their phone are the ones who actually come to you when something goes wrong. The device stays useful, and you stay in the loop.
The Bigger Picture
This study strikes a nerve with parents for a simple reason. It confirms a suspicion many have carried quietly, that the safety language on these apps functions partly as reassurance for adults rather than real protection for kids. Parents did the responsible thing, followed the instructions, and trusted the labels, and the research shows the labels often did not deliver.
The gap it exposes is one parents cannot close alone with better toggles. When the defaults are safe, kids are safer without anyone lifting a finger. When safety is optional, buried, and easy to circumvent, the burden shifts onto families and onto children themselves. The most useful thing the Northeastern team offers is not a longer checklist but a clearer standard for what real protection looks like: on by default, resilient to everyday use, and actually visible to the child it is meant to guard.
This is a sensitive topic, and it touches on self harm and disordered eating. If your child is struggling with these issues, a pediatrician or mental health professional can help you find the right support, and you do not have to sort it out on your own.