A federal appeals court has ruled that the First Amendment protects the private, in-home possession of AI-generated child sexual abuse images that depict no real child, even as the judge who wrote the decision pleaded with the Supreme Court to revisit the issue given how far artificial intelligence has advanced. The ruling came in the case of Steven Anderegg, a 42-year-old Holmen, Wisconsin resident who used a Stable Diffusion AI model to generate thousands of explicit images of children, and it leaves intact the more serious charges he still faces.
The U.S. Court of Appeals for the 7th Circuit upheld a decision from the Western District of Wisconsin, which had dismissed the possession charge last year, according to the Maryland Daily Record. The Justice Department originally charged Anderegg with producing, distributing and possessing obscene visual depictions of minors after bringing the case in 2024. Federal prosecutors appealed the dismissal of the possession count, but the appellate panel sided with Anderegg on that narrow question, even though he still faces pending charges for producing and distributing AI-created child sexual abuse material, and for allegedly transferring similar AI-generated sexual images to a 15-year-old boy via Instagram.
A Narrow Win Built on Old Precedent
Judge John Z. Lee, who wrote for the panel, pointed to Supreme Court rulings establishing that individuals have a right to possess obscenity privately in their homes, and separate rulings finding that virtual child sexual abuse material depicting no real person does not qualify as child pornography. The Supreme Court carved out one exception: morphing images of real children into sexual poses remains illegal because it harms actual children, a distinction the appeals court left untouched.
Lee acknowledged the discomfort built into that legal framework. “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,” he wrote, according to the same report. The judge was careful to note limits on his own ruling — it does not protect someone who downloads such images from a third-party site or takes a device storing them outside the home.
What the Ruling Does Not Change
Legal scholars quoted by the paper stressed that Anderegg's victory is narrow. Riana Pfefferkorn said the ruling will not stop federal prosecution of people who create virtual abuse material in the first place, since production and distribution charges remain untouched by the decision. Eric Goldman noted that virtual CSAM may be different from actual CSAM and could end up regulated, or constitutionally protected, differently depending on how future courts and legislatures treat it.
Daniel Lyons said tighter restrictions on AI-generated child sexual abuse material are necessary, and he added that the Supreme Court has a decent chance of taking up the case if prosecutors push it further. The tension Lyons describes sits at the center of a legal framework that has not caught up with generative AI: the federal child obscenity statute, 18 U.S.C. § 1466A, was enacted under the PROTECT Act of 2003 specifically to prosecute explicit depictions of minors who do not actually exist, after the Supreme Court's 2002 decision in Ashcroft v. Free Speech Coalition struck down earlier bans on virtual child pornography, according to the Department of Justice.
A Reporting Pipeline Under Strain
Online platforms have relied for decades on a reporting pipeline built around the National Center for Missing and Exploited Children, which works with law enforcement to confirm whether images or videos depict abuse of a real child by comparing suspected material against a database of digital fingerprints known as hashes. AI-generated synthetic images are straining that system, and NCMEC has told Congress that reports of suspected AI-generated abuse material are overwhelming the organization, per the Daily Record's reporting. NCMEC has also reported that tech-company tips often omit basic information law enforcement needs to act on them.
That strain shows up in the numbers. CyberTipline reports flagged for generative AI content jumped from 4,700 in 2023 to 67,000 in 2024, then surpassed 400,000 in 2025, according to data from MissingKids.org. The Stanford Center for Internet and Society has explained that detection systems rely heavily on hash databases built from known real victims, leaving AI-generated synthetic images difficult to flag automatically since generative models produce novel pixel patterns that do not match existing signatures, requiring resource-intensive manual review instead.
Meta, which owns Instagram, sent a cybertip to NCMEC about Anderegg's alleged activity, and a Stability AI spokesperson said the company that develops the Stable Diffusion software is committed to preventing AI misuse, including the creation and dissemination of CSAM. Congress has also moved to widen reporting obligations: the 2024 REPORT Act expanded mandatory reporting duties for electronic service providers to include online enticement and child sex trafficking alongside CSAM detection.
States and Congress Move Where Courts Cannot
While the 7th Circuit says it lacks authority to redraw Supreme Court doctrine, other branches of government have been filling gaps of their own. As of August 2026, 46 states have enacted legislation explicitly criminalizing AI-generated or computer-edited child sexual abuse material, with more than half of those laws passed just in 2024 and 2025, according to Enough Abuse. Congress passed the TAKE IT DOWN Act in May 2025 to penalize nonconsensual digital forgeries and sexually explicit AI content, and federal prosecutors in Ohio secured the first criminal conviction under that law in April 2026 after a defendant generated and published explicit AI imagery depicting both minors and adults.
Civil litigation is testing another front. Three Tennessee teenagers filed a federal class-action lawsuit in March 2026 against xAI, claiming its Grok chatbot was used to generate sexually explicit deepfakes of minors using public social media photos, according to the American Bar Association. Together, the state statutes, the TAKE IT DOWN Act and pending civil suits reflect a legislative and litigation push to curb synthetic abuse material even as the 7th Circuit's ruling stands as a reminder that constitutional doctrine written before generative AI existed still governs what prosecutors can charge — at least until the Supreme Court decides whether to weigh in.
