Protecting children in the digital era requires innovation, collaboration, and ethical leadership. By investing in responsible AI and strengthening cross-sector partnerships, we are working to ensure that technological progress advances — not threatens — children’s rights and safety.
When we convened our first roundtable on AI in 2019, we could not foresee how rapidly the field would evolve. But even then, we sensed what was coming: AI would become a critical tool in the fight against child sexual abuse — and at the same time, a technology that perpetrators would seek to exploit.
Investing early in uncharted areas is part of our DNA. Our five-year project Stella Polaris, funded by the Swedish Postcode Lottery in 2021, has enabled us to build deep expertise, pilot new tools, and bring together child protection actors, law enforcement, AI researchers, academics, and tech companies around a shared purpose. This open and collaborative approach has generated new partnerships and concrete results.
In 2025, AI advanced rapidly — reshaping both the risks children face and the opportunities to strengthen their protection. Large language models became more capable, multimodal features became widely used, and AI agents began to appear in everyday applications. At the same time, debates on AI governance and regulation intensified globally.
The scale of the challenge
The volume of child sexual abuse material online continues to grow at an alarming rate. In 2025, the Internet Watch Foundation received 312,030 reports of confirmed abuse material distributed globally. In Sweden, the number of reported documented sexual offences against children has increased from 906 cases in 2015 to over 6,000 in 2024. For law enforcement, reviewing this material is time-consuming and deeply harmful to the officers who do it.
AI supporting law enforcement
In June 2025, we supported a second Interpol DevOps Hackathon, bringing together the Swedish Police and international partners. During the week, they further developed an AI tool that generates text summaries of abuse material — reducing direct exposure while streamlining investigations. A quality assurance framework was established to ensure reliability and responsible use. The result: investigators can move faster, focus more effectively on identifying victims and perpetrators, and be better protected in the process.
“We’re grateful that Childhood recognized early what the field is only now acknowledging: sustainable child protection requires protecting investigators from the very content they’re fighting against. Childhood’s investment made that future possible. The Foundation gave us the resources and credibility to demonstrate that analyst wellness and investigative effectiveness go hand in hand, and that both are essential to keeping children safe.”
Sherrie Bosisto, Founder & Executive Director, Global Emancipation Network
AI tools in practice
In 2025, all five organizations within the Cluster Network Against Sexual Violence — a Swedish cross-sector collaboration focused on supporting young people exposed to or at risk of sexual violence — implemented Fronta, an AI tool developed with our support in 2023. Fronta enables cross-organizational data analysis of youth support chats, helping practitioners identify patterns and respond more effectively. We also began testing an adapted version of a Finnish chatbot, strengthened with expertise on sexual abuse and eating disorders, designed to provide young people with accessible, knowledge-based support.
Bringing global leaders together
A milestone in 2025 was the high-level workshop on Artificial Intelligence: Risks and Opportunities for Children, co-hosted with IADC and the Pontifical Academy of Sciences at the Vatican. The two-day event brought together experts from across sectors and regions to examine how emerging technologies may impact children’s safety and rights. The workshop was opened by Her Majesty the Queen, who participated throughout both days alongside Princess Madeleine — a powerful signal of the importance of this issue at the highest level.
What we have learned — and where we go next
Our early and sustained work on AI has confirmed two critical lessons: high-quality data and strong technical expertise are essential — but they must be combined with deep knowledge of children’s development, behavior, and rights. Only this integrated approach produces solutions that are ethical, relevant, and fit for purpose.
As AI continues to evolve at pace, so will our commitment. The question is no longer whether AI will shape child protection or not. The question is how we ensure it does so responsibly, equitably, and always with the child at the center.
“We saw Stella Polaris as a groundbreaking initiative. When the Postcode Lottery awarded the dream project grant in 2021, the use of artificial intelligence to combat child sexual abuse in Sweden was largely unexplored. The project addresses a clear and urgent need by establishing a national hub where child rights and AI come together. By bringing together researchers, tech actors, funders, and experts, it creates interdisciplinary momentum and concrete, innovative solutions in a long-neglected field.”
Helene Carlbark, Head of Charities, The Postcode Lottery
