Scope of this entry
This record covers category D at the level of detection, statute, charge and enforcement outcome only. The conduct is described in the sources linked below. It is not reproduced here, and nothing about how the material was produced is set out.
How the case surfaced
The editing platform flagged an account carrying explicit content and notified the National Center for Missing and Exploited Children. That referral is what reached Boulder County. The Boulder County Digital Forensics Lab took the investigation, and the sheriff’s office made the arrest on 30 January 2026, with bond set at $50,000.
It is worth being precise about the order of events, because it inverts the usual assumption. No police unit detected this. No victim reported it. A private company’s content moderation system produced a signal, a nonprofit clearinghouse routed it, and a criminal investigation began at the end of that chain rather than at the start.
Boulder County described this as among the first full investigations its forensics lab had run involving generative AI used to create this kind of material. The detection did not come from the lab. It came to the lab.
The statute
Every count is Colorado state law. Eleven counts of sexual exploitation of a child by producing sexually exploitative material under C.R.S. section 18-6-403(3)(b), a class 3 felony. Eleven counts of possession under section 18-6-403(3)(b.5), a class 5 felony. One count under section 18-6-403(5.5)(a), the extraordinary risk enhancer for material involving a child under twelve.
No federal charges. No statute written for artificial intelligence. The provisions used here predate generative image tools by a wide margin and do not mention them.
That is now the third case in this record where the law doing the work was not written for AI. It is worth holding alongside the first TAKE IT DOWN Act conviction, where the child-related conduct was also reached by an older statute rather than by the new one.
Altered, not invented
The sheriff’s account is specific on one point that most coverage treated as incidental. Photographs of real children were altered using generative tools. The material was not conjured from nothing.
That distinction carries legal weight in the United States. Wholly synthetic depictions occupy contested constitutional ground, which is why the federal provision covering them is framed around obscenity rather than treating the images as contraband outright. Material derived from identifiable real children does not raise that problem at all.
Whether prosecutors are consequently more willing to charge altered-image cases than wholly synthetic ones is a question this record cannot answer from a single case. It is flagged because it is testable, and because the distinction is doing more work in these prosecutions than the coverage suggests.
Open
No disposition. The arrest was in January 2026. This record has located no plea, trial date, dismissal or sentence in the eight months since. Charges are allegations and nothing here has been tested in court.
What nobody reported. No source states how many images were involved, how the photographs of real children were obtained, whether the children were identified, or whether their families were notified. Those are the questions that determine the actual harm in a case like this, and none of the coverage asked them.
The platform. No source addresses whether the flagging was automated or manual, how long the account operated before it was flagged, or whether the platform’s terms anticipated this use. A detection pathway that begins with private content moderation deserves more scrutiny than it has received.