The First Amendment protects an individual’s right to privately possess images or videos of child sexual abuse created using artificial intelligence — if the material does not depict a real person and remains in the home, a federal appeals court judge ruled Tuesday. The decision came in a case that tested the scope of laws implemented before advances in AI made it easy to create realistic-looking fake images of children.
Judge John Z. Lee, in his ruling for the U.S. Court of Appeals for the 7th District, urged the Supreme Court to reconsider the issue of virtual abuse material of children in the age of AI, suggesting that technological advances complicate existing legal precedent.
“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,” the judge wrote.
The Justice Department in 2024 brought charges against Steven Anderegg, of Holmen, Wisconsin, alleging that he used a popular AI image generator to create thousands of explicit images of children. He was accused of breaking federal obscenity laws by producing, distributing and possessing obscene visual depictions of minors engaged in sexually explicit conduct. Those laws can be applied to material not involving a real person.
Anderegg sought to have the possession and production charges dismissed under the First Amendment. Last year, a judge in the Western District of Wisconsin agreed to dismiss the possession charge, prompting the federal prosecutors to appeal.
Lee, in his ruling upholding the lower court’s decision, cited the precedent set by Supreme Court rulings from 1969 and 2002.
One held that individuals have a right to possess obscenity in the privacy of their home. The other found that “virtual” child sexual abuse material, also known as CSAM, not depicting a real person is not child pornography.
Tuesday’s decision came in a novel case in which the Justice Department is testing the limits of those obscenity laws to curb the growing problem of AI-generated child sexual abuse imagery.
Riana Pfefferkorn, a policy fellow at the Stanford Institute for Human-Centered AI, said the ruling won’t stop federal prosecution of people creating virtual abuse material, based on her research into related court cases this summer.
“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,” she said. Lee’s ruling does not protect acts like downloading images from a third-party site or taking a laptop or phone storing the images outside the home, Pfefferkorn said.
But the decision is “a reminder to the world that virtual CSAM may be different than actual CSAM,” said Eric Goldman, a law professor at Santa Clara University School of Law. “It might be regulated differently or protected by the Constitution differently.”
Daniel Lyons, a law professor at Boston College, said there is a decent chance the Supreme Court will take up the case. Last year, it revisited early-internet-era speech and pornography decisions that affect children in the course of upholding Texas’s right to require pornography websites to verify a user’s age before granting access.
Lyons said tightening restrictions on AI-generated child sex abuse material is necessary because the growing number of synthetic images is straining a system originally created to combat explicit imagery of real children, he said.
Online platforms have for decades relied on a reporting pipeline run by the nonprofit National Center for Missing and Exploited Children (NCMEC). The system allows companies to automatically compare suspected abuse material against the center’s database of digital fingerprints for known violating content, known as hashes.
NCMEC also works with law enforcement to confirm that an image or video depicts abuse of a real child. The nonprofit recently told Congress it is being overwhelmed by the deluge of reports from tech companies of suspected AI-generated abuse material and that those reports often leave out basic information that law enforcement needs to investigate.
Law enforcement learned about Anderegg’s activity through a “CyberTip” to NCMEC from Meta, which owns Instagram, after he allegedly sent images via direct message. Anderegg made images using an AI model released by Stable Diffusion, an open source developer.
A spokesperson for Stability AI, which develops the software, said in a statement that the company “is committed to preventing the misuse of AI technology, particularly in the creation and dissemination of harmful content, including CSAM.”
