Federal Appeals Court Rules Private Possession of AI-Generated Abuse Images Is Protected Speech

The legal boundary for artificial intelligence was not drawn by a modern tech regulator, but by a 2002 Supreme Court ruling on early computer graphics. That precedent established that an image cannot be treated as child sexual abuse material if no actual child was harmed to create it. Now, that distinction has forced a federal appeals court to dismiss a possession charge against a Wisconsin man who used a text-to-image model to generate thousands of hyper-realistic, sexually explicit pictures of fictional minors. On August 25, a three-judge panel of the U.S. Court of Appeals for the Seventh Circuit ruled that the First Amendment protects the private, in-home possession of entirely synthetic abuse imagery, provided no real person is depicted.[1][3]

The decision exposes a widening gap between twentieth-century constitutional law and the capabilities of modern generative AI. The defendant, 42-year-old Steven Anderegg, was arrested in May 2024 after an investigation that began when Meta flagged his Instagram account in October 2023. Federal prosecutors alleged that Anderegg, a software engineer, used the open-source model Stable Diffusion to manufacture the images, deliberately engineering the software to bypass its safety filters.[3]

Authorities seized multiple cellphones and a laptop from Anderegg's residence in Holmen, Wisconsin, uncovering a massive cache of synthetic material. He was subsequently indicted on four federal counts: producing, distributing, and possessing obscene visual depictions of minors, as well as transferring the material to a 15-year-old boy.[1]

Anderegg's defense team moved to dismiss the charges, arguing that the possession and production of entirely fictional images were protected speech. A federal district judge agreed to drop the possession count, prompting the Justice Department to appeal. The Seventh Circuit's ruling last week affirmed that lower court decision, leaving the government with a fractured case.[1][3]

The appellate panel's reasoning rests on a rigid application of two Supreme Court precedents. In the 1969 case Stanley v. Georgia, the Court ruled that the government cannot criminalize the private possession of obscene material inside a person's home. Thirty-three years later, in Ashcroft v. Free Speech Coalition, the justices struck down a federal ban on digital images that merely "appear to be" minors, concluding that because no actual child is exploited in the creation of a computer-generated image, the government lacks the justification to strip it of First Amendment protection.[2][4]

By combining those two rulings, the Seventh Circuit arrived at a strict constitutional math: entirely artificial images do not harm a real child during creation, qualifying them as protected virtual speech, and the government cannot punish a citizen for possessing protected speech in private. The ruling applies specifically to 18 U.S.C. § 1466A, a federal obscenity statute updated in 2003 to cover computer-generated images. The appeals court did not strike down the law entirely, but ruled it unconstitutional as applied to Anderegg's private possession.[3]

The judges who authored the unanimous decision made it clear they were reluctant to reach that conclusion. Writing for the panel, Judge John Z. Lee urged the Supreme Court to revisit the issue, noting that modern AI systems can produce output that is virtually indistinguishable from photographs of actual abuse. "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 wrote.[1][5]

Prosecutors had argued that the landscape has fundamentally changed since 2002. They contended that synthetic images are now so realistic that investigators are forced to expend limited resources verifying whether a real child is in danger, diverting attention from actual victims. The government also raised concerns that AI-generated material is actively used by predators to groom minors.[3]

The ruling does not grant blanket immunity to those generating synthetic abuse material. The Seventh Circuit's decision is narrowly tailored to private possession; Anderegg still faces three federal charges for producing, distributing, and transferring the images. The court also noted that the Supreme Court has previously carved out exceptions for "morphing"—editing the faces of real children onto explicit content—because that practice inflicts harm on an identifiable victim.[1]

The Justice Department has not yet indicated whether it will appeal the Seventh Circuit's decision to the Supreme Court. Until the high court agrees to reconsider its 2002 precedent, federal prosecutors face a legal framework where the distribution of synthetic abuse material is a felony, but hoarding it in a private residence is a protected constitutional right.[4]