Frontier AI Risks Initiative
Frontier AI - the most capable general-purpose AI systems being developed - could bring significant benefits, but it also presents credible risks of severe harm. These include the potential misuse of advanced capabilities for cyber attacks or biological threats, as well as concerns that increasingly capable systems could operate beyond effective human control. Children have a direct stake in how these risks are governed. They depend on adults, essential services, and public institutions for their safety and development, and will live with the consequences of decisions being made about AI today for decades to come.
In September 2026, international leaders launched A Call for Control of Frontier AI Models, calling for stronger company safety measures, coordinated government oversight, and international cooperation to ensure that the most advanced AI systems remain subject to human direction, oversight, and control. SAIFCA published Children’s futures and the governance of frontier AI in support of that call, with an additional focus on children. Our statement calls for children’s rights, safety, and long-term interests to be central to its implementation, for children’s rights and development expertise to inform standards and independent evaluations, and for children to have meaningful, age-appropriate opportunities to contribute to decisions affecting their futures.
Our Frontier AI Risks Initiative takes that work forward. SAIFCA aims to help connect specialist work on catastrophic AI risks and effective red lines with organisations and people working for children’s safety, rights, and futures. We are seeking stronger support for effective government and international action, while helping to ensure that children’s interests and relevant expertise are represented as governance develops. We do not seek to determine the technical thresholds for frontier AI risks, since that work requires specialist expertise in AI safety, security, evaluation, and international governance.
The resources below set out our public statement and recommendations, explain how catastrophic AI risks could affect children, examine the evidence behind our assessment, and look at the wider competitive pressures shaping advanced AI development. We will continue to develop this initiative through engagement with policymakers, children’s organisations, researchers, and experts working on frontier AI safety and governance.
Our Frontier AI Risks Initiative
SAIFCA
SAIFCA
SAIFCA
Understanding the Risks and the Case for Action
SAIFCA
SAIFCA
SAIFCA