Federal Appeals Court Rules AI Created CHILD PORN Is Protected Speech

A federal appeals court has just told prosecutors in Illinois, Indiana and Wisconsin they cannot charge a man for privately possessing AI-generated child sexual abuse images inside his home.
The images were found to contain only AI generated images of children. The court said Supreme Court precedent from 1969 and 2002 left it no other option. Production, distribution, and sending the material to a minor still stand.
The distinction is now already changing how ordinary people, such as professional photographers handle pictures of children.
The case is United States v. Steven Anderegg. Anderegg, of Holmen, Wisconsin, was charged in 2024 after Meta reported that AI-generated images had been sent to a minor on Instagram. Investigators later found hundreds of similar images on devices taken from his home.
Prosecutors say he used Stable Diffusion, a text-to-image program, plus extra software and carefully worded prompts, to generate photorealistic pictures of prepubescent minors.
The government conceded the images did not depict an actual child and could not be traced to one.
A Western District of Wisconsin judge dismissed only the federal possession count in 2025. The Justice Department appealed that single dismissal. On August 25, 2026, a Seventh Circuit panel of Judges John Z. Lee, Doris Pryor, and Joshua Kolar affirmed. The ruling applies in the Seventh Circuit only. It is not national law.
The panel did not bless the material. It did not wipe the case. It held that 18 U.S.C. § 1466A(b)(1) is unconstitutional as applied to Anderegg’s in-home possession of obscene, AI-generated images that do not depict a real child.
Two Supreme Court decisions controlled the outcome. Stanley v. Georgia (1969) protects private possession of obscene material inside the home. Ashcroft v. Free Speech Coalition (2002) held that sexually explicit images of fictional children, created without using real minors, do not fall under the narrower rules that apply when an actual child is exploited.
Because no real child appears in these pictures, the government’s usual justifications for banning possession were already rejected a quarter-century ago.
Judge Lee wrote the opinion and did not pretend the technology is the same as it was in 2002.
“We now live in an age where GenAI models can render images depicting the abuse of virtual children that are virtually indistinguishable from those depicting the abuse of actual children,” he wrote. “This case illuminates how this evolving technology complicates the lines drawn by the Supreme Court in Stanley, Osborne, and Free Speech Coalition.”
He added: “Indeed, in Free Speech Coalition, the Supreme Court addressed the scope of First Amendment protections for virtual CSAM, but that was nearly twenty-five years ago, and the image-generation technology available today was likely unimaginable back then. Given the relentless advancement in artificial intelligence models, we have some concerns about the lines these cases draw, but we are not free to redraw them ourselves.”
Lee and Kolar filed a concurrence asking the Supreme Court for new guidance if the right case arrives. In essence, lower courts just admitted the map is obsolete and said they cannot redraw it.
Anderegg still faces the production and distribution counts, plus the charge that he transferred the material to a boy under 16. Those are not protected by this ruling. He also faces separate state charges in La Crosse County, including first-degree child sexual assault and exposing a child to harmful material.
The legal line is narrow. The practical effect is not.
In Hereford, Texas, photographer Ashley Contreras began pulling children’s photos off her website and social media after the ruling. “AI has taken over and what you want it to be at one point, it’s just an innocent picture of somebody riding a bike completely different by the end of the day when falling into the wrong hands,” she said. “And it’s a lot harder to show examples of what I can do.” She also warned parents that deleting a photo from social media does not mean it is gone.
In Grand Rapids, Michigan, Skyla Davis of Skyla Davis Portraiture made the same call on her own. She will now post children only without identifiable faces and is deleting older public posts that show kids’ faces. Parents still get the full gallery. Public marketing does not.
“No one has reached out to me and asked me to do this. It was a decision that I made on my own,” Davis said. “As much as I want to say it was a tough decision, it wasn’t.”
“These children, they cannot consent, and we have a job to protect them as best as we can,” she said. “They’re so innocent. I mean, it hurts my heart because I look into my nephew’s eyes or my Godkids’ and I’m just like, they don’t have a clue in the world.”

“My biggest concern is that these policies are not keeping pace with the new age of AI,” she said. “My mind automatically thinks someone is looking at a picture that I post of someone’s child, saving it, and doing who knows what with it.”
That is the world the 2002 Court never saw. Cartoonish “virtual” images are no longer the problem. Photorealistic output from open-source models is.
The Seventh Circuit decision hit a federal possession statute. It did not strike Wisconsin Act 224, the 2024 state felony for receiving, distributing, producing, possessing, or accessing obscene material that depicts a “purported child” in sexually explicit conduct, whether the child is real or not. That law is still in effect. Lawyers are already lining up to test it.
State Sen. Jesse James, a Republican who authored the act and works as a part-time police officer, was blunt about the target. “We’re not talking about adult AI-generated porn. We’re talking about child AI-generated pornography. Just the idea of that in itself is absolutely disgusting.”
“Their whole concept is they want to target a child victim. That’s their bottom line,” James said. “We need to make sure that we’re keeping up with the speed of this.”
Anderegg’s attorney, Anthony Cotton, called the federal possession statute unconstitutional and said the same argument should take down the state law. “This is an unconstitutional statute … that should be struck down.”
Justice Clarence Thomas flagged the problem in 2002. “If technological advances thwart prosecution of ‘unlawful speech,’ the Government may well have a compelling interest in barring or otherwise regulating some narrow category of ‘lawful speech’ to enforce laws against material produced through the abuse of real children,” he wrote then. The technology he was warning about is now sitting on consumer laptops.
This is not a nationwide green light. It is a circuit-level holding that private possession of purely synthetic images, kept at home, cannot be prosecuted under one federal count when the government admits no real child was used. Create it, send it, or put it in front of a minor, and the case proceeds.
That is still a hole large enough for photographers to change how they work and for parents to treat every public photo of a child as raw material. The judges said they were bound. They also said the lines no longer match the machines. Until the Supreme Court or Congress updates those lines, families and small businesses are writing their own rules while the old ones sit on the books.
source modernity.news

Charles Higley
| #
Okay, I see why the holder of the child pron is not guilty of possessing child porn. However, how about preventing the source from producing the child porn? AI programs do not have any rights and such activities should be shut down and outlawed. AI should not be allowed to run amuck and do whatever sick minds want them to do. They should be limited in scope and heavily monitored.
Reply