Australia made headlines when it banned children under 16 from social media platforms. The UK followed with its own Online Safety Act provisions. Politicians in the United States have pushed similar legislation at both the state and federal level. On the surface, these moves look like meaningful action. According to IEEE Spectrum, researchers who actually study child online safety say the legislation is largely missing the point — and could even create new risks in the process. For a sense of how unchecked platform behavior can spiral into real-world harm, consider the OpenAI Medicare breach, where automated systems caused serious damage before anyone noticed.
The core problem is architectural. Social media platforms are not dangerous simply because children have accounts on them. They are dangerous because of how they are designed: recommendation algorithms that optimize for engagement above wellbeing, opaque content moderation systems, and data collection pipelines that build detailed behavioral profiles of minors. An age ban does nothing to change any of that infrastructure. A 15-year-old locked out of Instagram does not gain protection from the broader data economy, which continues to collect, infer, and monetize information about young users through dozens of other apps and services they are still permitted to use.
The Enforcement Gap Is Enormous
Age verification technology is nowhere near reliable enough to enforce these bans consistently. The methods platforms currently use — self-reported birthdates, credit card checks, government ID uploads — are either trivially bypassed or create serious privacy problems of their own. Requiring children to upload identity documents to a tech platform in order to prove they should not be on the platform is, to put it plainly, a strange way to protect them. Researchers cited in the IEEE Spectrum report note that stricter verification requirements could push minors toward less moderated corners of the internet, where there is no enforcement infrastructure at all.
There is also a disparity in how these bans affect different populations. Young people with fewer offline social resources — those in rural areas, LGBTQ+ youth who rely on online communities for peer support, or children with disabilities who use digital spaces for connection — bear a disproportionate cost when access is cut off bluntly. Blanket bans do not distinguish between a teenager doom-scrolling algorithmically amplified content and one maintaining a support network through an online community. The harm reduction calculus is far more complicated than legislators have publicly acknowledged.
What Would Actually Work
The IEEE Spectrum analysis points toward a different set of interventions: platform design regulation rather than access restriction. That means requiring companies to disable engagement-maximizing features — autoplay, infinite scroll, algorithmically curated feeds — for users under a certain age. It means mandating that default settings for minors be privacy-protective, rather than requiring families to opt out of data collection after the fact. It means enforceable data minimization rules that prevent platforms from building behavioral profiles on children at all, regardless of whether those children hold verified accounts.
Regulatory frameworks built around algorithmic accountability would be significantly harder to implement than an age ban, and significantly harder for platforms to lobby against with a simple compliance checkbox. But the research consensus is pointing in that direction. Age gates generate headlines. Structural platform reform generates safer environments. The distinction matters enormously, and right now, most of the legislative energy is going to the easier win. Policymakers serious about device-level vulnerabilities and digital safety should recognize that the same logic applies here: surface-level restrictions rarely address the underlying architecture driving the harm.
