IWF reports a rise in AI-made child sexual abuse images

The law already reaches AI

In the UK, child sexual abuse material is illegal whether or not AI was used to make it. So are modifying an AI model to produce it and passing such models to others. Developing a model intended to generate photorealistic material of this kind can carry a prison sentence of up to five years.

A government spokesperson said the law also prohibits AI models used to create the material, including shocking photorealistic images in which real children can often be recognised.

The policy question is therefore not whether the UK can prosecute this content under existing law. It is whether companies should face a new legal duty to build models that are harder to misuse.

A rise beyond the material itself

IWF analysts say AI-generated abuse material has increased over several years, with video growing especially quickly. The latest data, however, concern still images only.

The harm also reaches children whose likenesses are altered or fabricated. Report Remove, a service that seeks to remove intimate images from the internet, received 420 reports from children who said fake or altered images of them had been shared. That is already more than the 397 reports recorded across all of 2025.

The National Crime Agency and IWF have urged parents and carers not to post children’s photos on social media. Their advice also includes setting accounts to private or sharing pictures only with a “close friends” group.

The unanswered question is enforcement

UK AI minister Kanishka Narayan said “nothing is off the table” in discussions about new rules, but that proposed measures would be judged by whether they make people in Britain safer. IWF head of policy Hannah Swirsky backed calls from MPs for a mandatory AI law requiring companies to build safer models. She said technology companies must make tools that cannot be abused this way, and ensure their products are safe from the design stage.

I think the tension is between a law that already bans the material and a proposed duty to prevent its creation. The announcement offers no detail on how a new obligation would be measured or enforced. Without that, “safer models” is a direction, not yet a standard companies can be held to.