US protects private possession of AI-generated child sexual abuse images

• Court rules private possession of non-real-person AI abuse images is constitutionally protected.
• Judge Lee calls for reconsideration of laws governing AI-generated abuse material.
• Defendant still faces federal charges over producing and distributing AI abuse material.
• Rising synthetic imagery is overwhelming abuse-reporting systems.

The US First Amendment protects an individual’s right to privately possess child sexual abuse images created using artificial intelligence—provided the material does not depict a real person and remains in the home, a federal appeals court ruled Tuesday.

The ruling from the US Court of Appeals for the 7th Circuit came in a novel legal case testing the scope of federal obscenity laws enacted before advances in AI enabled the rapid creation of realistic fake imagery.

Writing for the panel, Judge John Z. Lee urged the Supreme Court to revisit existing legal precedents on virtual child abuse material, warning that rapid technological developments complicate historic judicial frameworks.

"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," Judge Lee wrote.

The case stems from 2024 charges brought by the Justice Department against Steven Anderegg of Holmen, Wisconsin.

Federal prosecutors alleged that Anderegg used a popular open-source AI model released by Stable Diffusion to generate thousands of explicit visual depictions of minors engaged in sexually explicit conduct.

Anderegg moved to dismiss his production and possession charges under the First Amendment.

Last year, a judge in the Western District of Wisconsin agreed to dismiss the possession charge, prompting an appeal by federal prosecutors.

In upholding the lower court's dismissal of the single possession count, Judge Lee cited Supreme Court rulings from 1969 and 2002.

One precedent established that individuals have a constitutional right to privately possess obscenity inside their homes, while the other determined that "virtual" child sexual abuse material (CSAM) not depicting an actual person does not legally constitute child pornography.

The ruling applies exclusively to purely virtual, uncirculated images.

The decision noted that the Supreme Court already established an exception for "morphing" images of real children into sexual poses, which remains illegal as child pornography because it causes direct harm to actual children.

Anderegg’s attorney did not immediately respond to requests for comment.

His charges for producing and distributing AI-created abuse material remain pending, alongside a criminal charge for allegedly transferring AI-generated sexual images to a 15-year-old boy via Instagram.

Legal scholars noted that the ruling offers only narrow protection for private possession. Riana Pfefferkorn, a policy fellow at the Stanford Institute for Human-Centered AI, emphasized that the decision will not halt federal prosecutions for creating virtual material.

"If all you can do is privately possess something, well, how do you come into possession of it? Either you produce it, or you receive it from somewhere," Pfefferkorn noted.

He added that acts like downloading images from third-party sites or moving devices storing the images outside the home remain unprotected.

Eric Goldman, a law professor at Santa Clara University School of Law, described the decision as "a reminder to the world that virtual CSAM may be different than actual CSAM" and may face different constitutional or regulatory standards.

Daniel Lyons, a law professor at Boston College, indicated a strong likelihood that the Supreme Court will eventually take up the issue.

Lyons stressed that tighter restrictions are becoming necessary because the influx of synthetic material is straining law enforcement and reporting pipelines.

Online platforms rely heavily on a reporting pipeline operated by the non-profit National Center for Missing and Exploited Children (NCMEC) to match digital fingerprints, or hashes, against known abuse content.

NCMEC recently informed Congress that it is being overwhelmed by a flood of tech company reports regarding suspected AI-generated abuse material, which frequently omit crucial details required for criminal investigations.

Law enforcement originally uncovered Anderegg’s activities after Meta reported a direct message tip to NCMEC from Instagram.

In a statement, a spokesperson for Stability AI affirmed that the company "is committed to preventing the misuse of AI technology, particularly in the creation and dissemination of harmful content, including CSAM."