The Big Problem: Safety Features That Pretend to Protect Kids
Social media child safety features are tools built into platforms like Instagram, TikTok, Snapchat and YouTube that claim to protect minors from harmful content and unwanted adult contact, yet new research shows that at least half of these social media safety features are either broken, buried, or missing, meaning they often fail to deliver the child protection online that families are led to expect. The key takeaway is blunt: parents should treat platform-built protections as marketing promises, not reliable safeguards. Researchers from New York University and Northeastern University tested 86 child safety features across those four apps and found around 60 percent failed at their intended purpose. Out of those 86 features, only 35 worked as described; the rest counted as child safety features failing because they did not function, were easy to bypass, or were hidden so deeply in settings that real-world use was unlikely.
What the Research Shows About TikTok, Instagram, Snapchat and YouTube
The study’s most damning finding is that failures were “critical and pervasive” across TikTok, Instagram, Snapchat and YouTube. Researchers created fake child and adult accounts and tested three scenarios: a child using the app normally, a teen trying to bypass restrictions, and a malicious adult attempting to reach a teen. The numbers tell the story of TikTok Instagram safety breakdowns: Snapchat had a 73 percent failure rate, Instagram 66 percent, YouTube 55 percent and TikTok 50 percent. One quotable conclusion from the report is that “on average, around 60 per cent of the features failed at their intended purpose.” These child safety features failing ranged from promised content blocks that did not work to contact limits that vanished the moment a child took the first step toward an adult account.
Real-World Risks: Harmful Content and Unrestricted Adult Contact
The practical impact on ordinary users is not abstract—it is about harmful content and direct adult access to minors. On Snapchat, the researchers’ adult account could “search for, find and then message the child account with zero restrictions,” making a mockery of child protection online. TikTok, which markets itself heavily on safety, recommended searches tied to eating disorders and self-harm, including phrases like “how to pretend to eat your food” and “razor blade skin” when teens looked for such topics. Instagram blocked some adult-initiated contact, yet if a child messaged an adult first, the adult could respond freely. Across all four platforms, the study found that promised blocks on harmful searches did not work as advertised. In practice, this means children can still find self-harm and pro-anorexia communities despite supposed protections designed to stop them.
Why Trust in Platform Safety Is Crumbling
Parents and regulators have been told for years that social media safety features are getting better, yet this research shows a gap between what platforms sell and what children experience. Companies contest the findings, arguing the tests do not reflect real children’s behavior and accusing the study of vague claims that misrepresent features. But another outlet was able to replicate the results, strengthening the case that this is not a lab artifact but a persistent failure. At the same time, platforms face lawsuits from school districts and individuals who say social media caused harm, and some governments are debating or implementing youth social media bans. Researchers warn that companies keep answering each crisis with “safety theatre”—announced tools that sound impressive, fail in practice, and are quietly forgotten until the next scandal. That pattern erodes any reasonable confidence in platform-led child protection online.
What Parents Should Do While Platforms Play Safety Theatre
This study does not mean parents must ban every app, but it does demand a change in strategy. With child safety features failing across major platforms, families should assume built-in tools are partial aids, not a complete shield. Relying on social media safety features alone—whether content filters or default messaging limits—is no longer a defensible approach. Instead, parents need to treat these platforms as hostile environments that may surface self-harm, eating disorder content, and unwanted adult contact despite official rules. The researchers argue that companies must move beyond that recurring cycle of critical failures patched by fresh promises and instead design products that are safe by default, with tools that are functional and easy to find. Until that happens, the safest assumption for TikTok Instagram safety and beyond is that if a feature sounds too protective to be true, today’s evidence suggests it probably is.
