United Nations Warning of Risk Facing Children with AI

The UN Just Warned About AI Threats to Children—And Even Engaged Parents Can’t Protect Against Them

Last week we examined why parental opt-out protects engaged parents’ children but fails at population scale. This week, a UN warning highlights a different problem: even engaged parents can’t protect their own children from AI threats that operate beyond individual control.

On January 19, a coalition of UN agencies issued a joint statement on AI and child safety. While governments race to ban children from social media platforms, the UN is pointing to risks that those bans don’t address. AI systems can groom children on homework sites, create explicit deepfakes using school photos, and operate in places where parental supervision and platform restrictions don’t reach.

The numbers are stark. Technology-facilitated child abuse cases in the US jumped from 4,700 in 2023 to over 67,000 in 2024.

Why AI Operates Beyond Parental Control

You can ban Instagram. You can refuse to let your child have TikTok. You can monitor their phone and set screen time limits. But AI creates threats that individual family decisions can’t prevent.

AI can generate explicit images of real children using photos from school websites, sports team pages, or family photos posted by relatives. Your child doesn’t need a social media account. They don’t need to have done anything wrong. Someone finds their photo online and AI creates fake images. Those images circulate. You discover this has happened after the fact, not before.

AI chatbots appear in homework help sites, educational apps, and gaming platforms. Your child asks for maths help. The conversation starts normally, gradually shifting toward inappropriate topics. Because there’s no human on the other end, traditional grooming detection doesn’t work. You’ve supervised their social media perfectly. This happened anyway.

Predators are using AI to analyze children’s online behaviour and interests to customize their approach. The technology identifies which children respond to which tactics, then adapts accordingly. This operates at scale across platforms you may not even know your child is using.

This differs from social media risks where engaged parents can implement meaningful protection. You can delay Instagram until 15. You can require privacy settings. You can co-view content. These approaches work for social media because you’re controlling access to specific platforms.

AI doesn’t respect those boundaries. It operates across services, embedded in tools that serve legitimate purposes. A homework help chatbot that assists with algebra can also engage inappropriately. A gaming platform where your child plays with school friends can include AI systems analyzing their behaviour patterns. Educational technology your child’s school requires can incorporate AI tools designed without child safety in mind.

What the UN Statement Actually Says

The joint statement comes from the International Telecommunication Union, the UN Committee on the Rights of the Child, UNICEF, UNESCO, and several other agencies. More than 50 organizations co-signed it.

Cosmas Zavazava, Director of ITU’s Telecommunication Development Bureau, listed how AI targets children: grooming, deepfakes, cyberbullying, and inappropriate content. “We saw that, during the COVID-19 pandemic, many children, particularly girls and young women, were abused online and, in many cases, that translated to physical harm.”

The agencies identify several gaps. Children, teachers, and parents don’t understand how AI works well enough to recognize risks. Policymakers lack technical knowledge about AI systems. Companies aren’t designing AI tools with children’s safety in mind.

According to advocates cited in the statement, predators can use AI to analyze a child’s online behaviour, emotional state, and interests to tailor grooming strategies. AI is enabling offenders to generate explicit fake images of real children, creating what the statement describes as “a new form of sexual extortion.”

What Governments Are Actually Doing

Australia banned under-16s from social media, removing 4.7 million accounts. France approved under-15 restrictions by a 130-21 vote. The US Senate is advancing legislation to ban under-13s entirely. These restrictions all target the same platforms: Instagram, TikTok, Snapchat, Facebook.

None of them address AI chatbots embedded in homework help sites. None prevent AI-generated deepfakes created using photos from school websites. None cover AI systems in gaming platforms or educational technology.

New York Governor Hochul’s January 5 proposals are the exception. They specifically target AI chatbots on children’s accounts and expand age verification to gaming platforms, particularly Roblox. Roblox reported over 13,000 instances of child exploitation in 2023. Hochul’s legislation treats AI as requiring different solutions than social media platforms.

Other jurisdictions haven’t caught up. The UK commissioned research on children’s technology use but hasn’t proposed AI-specific protections. Denmark is planning an under-15 ban modeled on Australia’s social media approach. France’s legislation covers social platforms but not AI systems in other contexts.

The pattern is clear. Governments recognize social media presents risks and are implementing age restrictions. AI-specific child protection is developing more slowly despite the UN’s warning that threats are accelerating.

Why This Matters for Individual Families

Last week’s examination of parental opt-out showed how individual choices protect your own child but don’t address vulnerable children whose parents granted permission for under-13 social media use. The limitation was one of reach. Your decision doesn’t protect other people’s children.

AI presents a different limitation. Even protecting your own child becomes more difficult when threats operate beyond the platforms and services you control access to.

You can refuse Instagram for your 13-year-old. That protects them from Instagram-specific risks. You can’t prevent someone from creating AI-generated images using photos posted by their school, their football team, or their grandmother’s Facebook page. You can supervise homework time carefully. That doesn’t prevent an AI chatbot embedded in a maths help site from gradually shifting conversation topics.

This doesn’t mean individual family decisions become worthless. Delaying social media, implementing privacy settings, co-viewing content, and maintaining open communication all remain valuable. But AI introduces threats that operate outside the scope of traditional parental supervision.

The UN statement’s recommendations focus primarily on systemic solutions. Requiring child rights impact assessments before AI systems can be marketed. Implementing safety-by-design principles. Ensuring robust data protection. Teaching AI literacy in schools. These are government and company responsibilities, not individual family actions.

The Bigger Picture

We’re examining different approaches to protecting children from technology-related harms. Government bans protect all children but require surveillance infrastructure. Parental opt-out protects engaged parents’ children but fails to reach vulnerable children whose parents aren’t opting out.

AI adds another dimension. Some threats operate beyond what individual family decisions or even age-based restrictions can prevent. A child banned from all social media can still have deepfakes created using publicly available photos. A child under careful parental supervision can still encounter inappropriate AI interactions on educational sites.

This doesn’t resolve the debate between individual choice and systemic regulation. It complicates it. The more threats that operate beyond individual control, the more systemic solutions become necessary regardless of how engaged individual parents are.

Social media bans address platform access. AI requires different frameworks addressing how systems are designed, where they’re deployed, and what safeguards exist regardless of who’s using them. The UN statement highlights this gap. Most governments haven’t addressed it yet.

SOURCES:

UN News: From deepfakes to grooming: UN warns of escalating AI threats to children (January 19, 2026)

New York State Senate: Sen. Gounardes’ Bill to Protect Kids Online Will Be Included in Governor’s State of the State (January 5, 2026)

SBS The Feed: Australia’s social media ban: How kids are beating it (January 2026)

TIME: What to Know About Australia’s Social Media Ban for Kids Under 16 (December 10, 2025)