UNICEF reports that children are embracing AI at a rate three times quicker than that of adults.

      A study involving 10 countries reveals that 20 million children are already using AI tools, while governance struggles to keep up with a generation that UNICEF refers to as being part of "a global experiment." According to a UNICEF statement released on June 30, these children are adopting AI technology more than three times faster than the adults surrounding them.

      This statistic comes from a new analysis released just before the inaugural Global Dialogue on AI Governance, accompanied by a caution: the regulations intended to safeguard children online are failing to adapt to the rapid pace at which they are engaging with AI.

      The data is part of the Disrupting Harm Phase 2 initiative, a research project conducted by UNICEF’s Office of Strategy and Evidence at Innocenti in collaboration with ECPAT International and INTERPOL, with funding from Safe Online.

      The fieldwork encompassed Armenia, Brazil, Colombia, the Dominican Republic, Jordan, Mexico, Montenegro, North Macedonia, Pakistan, and Serbia, with approximately 1,000 internet-savvy children aged 12 to 17 and 1,000 of their parents or caregivers surveyed in each nation.

      UNICEF and Ipsos then adjusted the national figures using UN population data to develop global estimates. Two key findings emerged from the release.

      Over two million children, roughly one in ten, reported seeking AI for guidance on issues that concern them. Another estimate indicates that 13 million children utilize AI tools for assistance with school assignments and homework, representing a more typical yet significantly larger application.

      UNICEF’s perspective on the matter is not one of celebration. The organization stated, “Children are increasingly exposed to AI systems, including their designs, underlying business models, and the usage of their personal data, yet they possess far less ability to evade or contest these technologies.”

      The report highlights that children are the first to experience the effects of inadequate governance, suffering the longest from its consequences, while most AI governance approaches do not regard them as a separate demographic.

      This disparity between rapid adoption and insufficient protection has already influenced discussions regarding online safety legislation for children currently underway in the US, as well as Florida’s lawsuit against OpenAI concerning chatbot safety for younger users.

      The children surveyed are not oblivious to the potential dangers; UNICEF’s data shows that one-third expressed concerns regarding AI being used for scams or misinformation. Additionally, a quarter voiced fears about their images or videos being altered into sexually explicit deepfakes, a worry previously mentioned by UNICEF in a statement about deepfake abuse linked to the same research initiative.

      UNICEF's recommendations for governments and the private sector are encapsulated in a five-point plan. It advocates for increased research on AI’s impact on child development, stricter legislation against AI-facilitated sexual exploitation, built-in safety and transparency in AI systems, expanded AI literacy support for children and caregivers, and investment in connectivity to prevent further disparities among countries.

      While not new for the organization, the magnitude of the adoption figure underlying these recommendations is notable. The figure indicating three times the adoption speed refers to the rate of uptake rather than the amount of use, a nuance that UNICEF's own language does not fully clarify. Many adults are still navigating the use of generative tools, a trend monitored by TNW within workplace adoption statistics.

      Initiatives like Malta’s national AI literacy program suggest one potential policy approach, combining access with structured education before children and parents are left to figure things out independently.

      UNICEF has not specified a timeline for when its suggestions might be implemented, nor identified which governments or companies it believes are lagging the most. However, it clearly states that the opportunity to shape regulatory frameworks is diminishing at the same pace that children are engaging with these applications.