What happened

James Castell, 40, of Heathfield, East Sussex, used AI to create thousands of indecent images of minors, some as young as three years old.

Upon searching his home, officers from Sussex Police's Online Child Abuse Team (OCAT) seized digital devices containing over 3,800 indecent images, including 640 classified as Category A (the most serious level involving penetrative or sadistic acts).

Investigators found specialised AI software designed to generate images from text prompts, alongside evidence that Castell had been searching for and producing material involving children. 

Castell pleaded guilty to multiple charges and was sentenced to 18 months in prison (suspended for two years), placed on the UK Sex Offenders Register for ten years, and handed a 10-year Sexual Harm Prevention Order (SHPO) to restrict his digital access.

Why it happened

The incident was driven by the increasing accessibility of open-source or illicitly modified (such as Stable Diffusion) generative AI tools that allow users to bypass traditional creative barriers to produce hyper-realistic "pseudo-photographs."

While many mainstream AI developers (like Google or OpenAI) implement strict "safety filters," the rise of unfiltered, locally-run AI software means bad actors can generate harmful content without corporate oversight.

The "Pseudo-Photograph" Gap: Historically, legal frameworks struggled to categorize AI-generated content. This case was part of a broader effort by UK prosecutors to prove that AI-generated imagery meets the legal threshold of "pseudo-photographs," ensuring offenders cannot hide behind the "synthetic" nature of the media.