What happened
The Washington Post reported that the 7th Circuit upheld dismissal of a possession charge against Wisconsin defendant Steven Anderegg, who prosecutors accused of creating thousands of AI-generated explicit images of minors. Judge John Z. Lee said existing Supreme Court precedent may protect private possession of wholly virtual material kept in the home, while urging the Supreme Court to reconsider how those precedents apply as AI image generation advances.
The Washington Post, in the report republished by Minnesota Lawyer, said the 7th Circuit issued a decision Tuesday in the federal prosecution of Steven Anderegg of Holmen, Wisconsin. The Justice Department charged Anderegg in 2024, alleging he used an AI image generator to create thousands of explicit images of children. Charges included producing, distributing and possessing obscene visual depictions of minors engaged in sexually explicit conduct. According to the report, federal obscenity laws can reach material not involving a real person, the legal basis prosecutors used to pursue the case.
The report said a Western District of Wisconsin judge dismissed the possession charge last year after Anderegg challenged it under the First Amendment. Federal prosecutors appealed, but the 7th Circuit upheld that decision. Judge John Z. Lee wrote that private possession of AI-generated images may be constitutionally protected when they do not depict a real person and remain in the home. The Washington Post reported that Lee cited Supreme Court decisions from 1969 and 2002 concerning private possession of obscenity and virtual child sexual abuse material. Lee also wrote that rapidly improving AI complicates those precedents, but the court could not rewrite the legal boundaries itself.
The ruling did not resolve the rest of the prosecution. The Washington Post reported that production and distribution charges remain pending, as does a separate allegation that Anderegg transferred similar AI-generated images to a 15-year-old boy through Instagram. Law enforcement learned about the alleged activity through a CyberTip submitted to the National Center for Missing and Exploited Children by Meta, which owns Instagram. It also said Anderegg used an image model identified as released by Stable Diffusion. The source package does not include the appellate opinion, docket materials or an independent account from the Justice Department or Anderegg’s lawyer, so the ruling’s exact wording and procedural posture are not independently confirmed here beyond the Washington Post report.
Source details: minnlawyer.com ↗
Why it matters
The ruling draws a legally important distinction between synthetic images and material depicting a real child, but the source does not establish that creating, distributing, downloading or transferring such material is broadly lawful. The Washington Post reported that production and distribution charges against Anderegg, along with an allegation that he sent images to a 15-year-old, remain pending.
The case matters because it tests whether laws written around real-world abuse imagery apply in the same way to wholly synthetic material. The Washington Post reported that the Supreme Court has treated virtual images differently from images involving real children, while recognizing an exception for “morphing” images that use a real child’s face or likeness in a sexualized depiction. The distinction raises whether constitutional analysis turns on apparent appearance, harm to an identifiable child, conduct used to create or share it, or some combination.
The reported holding is narrow but consequential. Riana Pfefferkorn, a policy fellow at Stanford’s Institute for Human-Centered AI, told the Washington Post that the decision would not prevent federal prosecution of people who create virtual abuse material. She also said the ruling does not protect downloading images from a third-party site or carrying a laptop or phone containing the images outside the home. Eric Goldman, a Santa Clara University law professor, described the decision as a reminder that virtual and actual CSAM may be regulated or constitutionally protected differently. These are expert interpretations reported by the outlet, not an independently verified description of every possible application of the opinion.
The case also intersects with the operational burden on platforms and investigators. The Washington Post reported that NCMEC recently told Congress it was overwhelmed by a surge of reports from technology companies involving suspected AI-generated abuse material, and that some reports lack basic information needed for investigation. Traditional hash-matching systems identify known files, while synthetic material may not match an existing database record. That does not determine the legal question, but it shows how generative tools can affect reporting pipelines, triage and investigative resources while courts define criminal-liability boundaries.
What to watch next
Key unknowns include the precise scope of the appellate opinion, whether prosecutors seek further review, and whether the Supreme Court takes up the case. The Washington Post also reported that NCMEC is being overwhelmed by reports involving suspected AI-generated abuse material, creating a separate challenge for platforms and investigators.
The first issue to watch is whether the Justice Department seeks further review. The Washington Post reported that Boston College law professor Daniel Lyons believed there was a meaningful chance the Supreme Court would take the case, partly because the dispute concerns how older speech and obscenity precedents apply to modern AI-generated material. That assessment is an expert prediction, not evidence that the Supreme Court has agreed to hear the case. The source does not report a petition, a Supreme Court order or a timetable for further proceedings.
The second issue is how lower courts distinguish private possession from conduct that commonly precedes it. The reported decision concerns material kept in the home, while the prosecution’s remaining counts concern alleged production, distribution and transfer. The source does not provide the full opinion’s treatment of caching, cloud storage, automated generation, temporary files, third-party services or images derived from real children’s likenesses. Those details could determine how widely litigants apply the ruling.
The third issue is the policy response to synthetic abuse imagery. The Washington Post reported that Lyons argued tighter restrictions are needed because increasing volumes of synthetic images are straining systems created to address abuse involving real children. Stability AI told the outlet that it is committed to preventing misuse of its technology, including the creation and dissemination of CSAM. The source does not independently establish the effectiveness of that company’s safeguards, the scale of any particular platform’s synthetic-content problem, or whether Congress and the Supreme Court will change the law. Those remain meaningful unknowns.
