AI-generated child sexual abuse material is a rising threat. Artificial intelligence can now generate hyper-realistic child sexual abuse images in the absence of a real child, dismantling long-standing assumptions about victims, harm and criminal law.
Before AI, the criminalisation of child sexual abuse material was straightforward. Any sexual abuse image involving a real-life child could be easily criminalised on the basis that it included the exploitation of a real child. AI-generated child sexual abuse material has complicated this framework.
If a real child hasn’t been exploited, there is a loophole in the once rigid system. This leaves the courts to grapple with the difficult question of how to punish the creation of such material when there is no real victim involved?
The traditional legal framework
Existing legislation does cover AI-generated child sexual abuse material. Section 1 of the Protection of Children Act 1978 prohibits creating, distributing or possessing with intent to distribute an indecent photograph or pseudo-photograph of a child. The Act defines a pseudo-photograph as “an image, whether made by computer graphics or otherwise, which appears to be a photograph.”
This means it’s irrelevant whether the material is an actual photograph of an actual child or an image that has been created by technology and/or AI and/or other means if it appears to be a photograph.
Similarly, while photorealistic AI images are likely to fall within the definition of a pseudo-photograph, more obviously AI generated images could well amount to prohibited images under the Coroners and Justice Act 2009. As generative AI produces images that increasingly blur the line between synthetic and photographic, uncertainty could arise over which offence applies. This means recent reforms have focused less on closing an obvious legal loophole than responding to the scale and sophistication of AI-generated abuse material.
Addressing the non-existent child in AI-generated child sexual abuse material
The law currently makes little distinction between a genuine photograph of a child, an AI‑generated sexual image based on a real child and a synthetic image of a child who does not exist.
Sentencing tends to follow this same approach, though there is a strong argument that it should not. A meaningful distinction is needed, particularly between AI‑generated sexual images and real photographs, because the level and nature of harm differ significantly. Genuine photographs almost inevitably involve the abuse of an actual child, while AI‑generated images depicting identifiable children also carry a substantial risk of harm. Entirely synthetic images, however, present more complex questions, as no direct victim exists and the harm becomes more abstract.
The UK responds to AI-generated child sexual abuse material
Parliament has recently expanded the law beyond the images themselves to target the AI systems that facilitates their creation and distribution. Â The Crime and Policing Act 2026 introduces offences covering AI tools designed to generate child sexual abuse material, the possession and distribution of manuals explaining how to create such images using AI.
The rapid evolution of generative AI has forced lawmakers to reconsider what these offences are intended to protect. Rather than relying solely on traditional concepts of victimhood, the law is increasingly focused on preventing technological exploitation and protecting children from emerging forms of abuse and safeguarding the public from the risks posed by AI-generated child sexual abuse material.
If this blog touches on a case you are working on and you’d like to discuss this with one of our experienced criminal barristers, please contact our clerks today.