Thorn collaborates with UK AI Security Institute to release a safety protocol for generative AI

At Thorn, we believe technology can and must protect children, not put them at risk. That’s why we’re proud to announce a new milestone in our global child protection work. We’ve co-authored a safety protocol for AI-generated child sexual exploitation and abuse (CSEA) with the UK AI Security Institute (AISI).

This safety protocol creates a framework to prevent AI-generated CSAM. It considers the entire AI lifecycle, from model training and deployment to content hosting and moderation. This is a roadmap for developers, model hosting services, and others to build AI systems with child safety at their core.

Collaborations like this represent a significant step forward in Thorn’s Safety by Design initiative, first launched with All Tech Is Human, to guide ethical AI development. The new safety protocol expands that foundation, aligning with ongoing standardization efforts. It brings Thorn’s expertise into direct partnership with the UK AI Security Institute, which is the world’s biggest state-backed organisation of its kind.

Why this matters

The misuse of AI to further child sexual exploitation and abuse is an emerging and serious threat. Without clear safeguards, these technologies can be exploited to create abuse material. By establishing prevention frameworks now, this safety protocol helps stop these harms at scale.

How this fits into Thorn’s broader strategy

This work is one part of Thorn’s comprehensive AI safety strategy, built around three pillars:

• Transparency and improvement. Sharing transparency reports that elevate companies’ commitments and encourage shared responsibility.
• Standards and scalability. Collaborating with global standard-setting bodies to make safety interventions measurable and auditable.
• Policy and partnership. Engaging with policymakers to ensure emerging legislation is both feasible and impactful.

Each of these pillars helps make safety a default in every layer of our digital world.

Looking ahead

Technical standards are among the most powerful tools we have to keep children safe. They ensure consistent practices across industries, enable third-party audits, and create shared accountability. With this safety protocol, Thorn is growing the global reach of the existing Safety by Design principles and mitigations while adding new practices to address the evolving nature of generative AI-enabled sexual harms against children.

This new safety protocol represents an impactful step toward global responsibility of AI systems. It strengthens the safety infrastructure to prevent technology from harming children and reaffirms Thorn’s role as a trusted voice in technology, policy, and child protection.

With the UK AISI, we’re proving that proactive partnership can help make the world safer for every child.