By Jessica Schidlow, Esq., Senior Policy Analyst & Legislative Counsel, ENOUGH ABUSE

Generative artificial intelligence can now produce virtual sexual depictions of children (AI-generated child sexual abuse material, or “AI-CSAM”) that are nearly indistinguishable from images documenting the abuse of actual children. A recent federal appellate court decision has brought into sharp focus the tension between that technology and a legal framework developed long before such imagery was possible.

The headlines were alarming, but the AI-CSAM ruling is considerably narrower than they suggest—and it does not make AI-CSAM broadly legal.

The AI-CSAM Ruling: What Was Before the Court

In the recent AI-CSAM ruling, United States v. Anderegg, federal prosecutors charged a Wisconsin man with producing, distributing, possessing, and transferring sexually explicit AI-generated depictions of children. According to the government, Anderegg used Stable Diffusion, a generative-AI program, to create more than 13,000 realistic sexual images of prepubescent children. Critically, the government could not tie a single one of the 13,000 charged images to an actual child (images that appear to depict identifiable children are known as sexual deepfakes).

The investigation began after Anderegg sent one of the images to a 15-year-old boy on Instagram. Meta reported the communication to the National Center for Missing & Exploited Children (NCMEC), ultimately leading investigators to Anderegg and his devices.

In February 2025, a federal district court dismissed only the possession charge. The court held that 18 U.S.C. § 1466A(b)(1), which prohibits possession of certain obscene virtual depictions of children, was unconstitutional as applied to Anderegg’s private possession of wholly virtual material in his home. The government appealed.

On August 25, 2026, the U.S. Court of Appeals for the Seventh Circuit upheld that dismissal. Because the ruling was an as-applied decision, the court did not invalidate the statute in every circumstance. It left intact the charges alleging production, distribution, and transfer of the material to a minor. Anderegg still faces those charges. That narrow posture is essential to understanding what the decision does—and does not—mean for child safety.

Why the Court Reached That Result in the AI-CSAM Ruling

The Seventh Circuit concluded that it was bound by several older United States Supreme Court cases.

In Stanley v. Georgia (1969), the Supreme Court held that although obscenity generally falls outside First Amendment protection, the government may not punish a person merely for possessing obscene material in the privacy of the home.

The Court later recognized a fundamentally different rule for CSAM depicting real children. In New York v. Ferber and Osborne v. Ohio, the Court emphasized that CSAM is distinct from ordinary obscenity because its production requires the exploitation of an actual child and its continued possession and circulation perpetuate that victim’s harm. The government therefore may criminalize even private possession of CSAM depicting actual children.

Then, in 2002, the Supreme Court decided Ashcroft v. Free Speech Coalition. There, the Court struck down portions of a federal statute that extended child-pornography prohibitions to “virtual” images created without the use of real children. Without an actual child victimized in producing the material, the Court held, the government could not simply apply the rationale that justified prohibiting CSAM involving real children.

That precedent forced the Seventh Circuit to draw a difficult line: material involving an actual child (deepfakes) may be prohibited even in the privacy of the home, while wholly synthetic imagery is entitled to substantially greater constitutional protection. The panel openly questioned whether that distinction still makes sense in the age of generative AI, and two of its members wrote separately to say so—a concurrence discussed below.

What the AI-CSAM Ruing in the Anderegg Case Does—and Does Not—Mean

The most important point is also the easiest to lose in the headlines: the AI-CSAM ruling in the Anderegg case did not hold that AI-generated CSAM is generally protected by the First Amendment.

It creates no constitutional right to produce, distribute, sell, transmit, or send obscene AI-generated sexual depictions of children. Its holding concerns only the application of a federal possession statute to wholly synthetic, obscene material privately possessed in the home, where the government could identify no actual child depicted in or victimized through the creation of the charged images.

It is equally important to distinguish wholly synthetic imagery from AI-generated or manipulated sexual images of real, identifiable children (deepfakes).

In Free Speech Coalition, the Supreme Court expressly distinguished so-called “morphed” images—images created by altering a real child’s photograph to make that child appear to be engaged in sexual conduct. Deepfakes such as those implicate the rights and interests of an actual child and therefore stand on very different constitutional footing. Federal law has treated such material as CSAM for decades. That distinction matters enormously for families.

If someone takes a photograph of a real child from social media and uses AI to create a sexual deepfake of that child, the AI-CSAM ruling in the Anderegg case does not extend constitutional protection to that conduct.

Federal law also separately prohibits publication of certain non-consensual AI-generated intimate imagery involving identifiable people, including minors, through the TAKE IT DOWN Act.

Why AI Changes the Stakes

The Supreme Court precedents shaping Anderegg date from 1969, 1990, and 2002. Generative AI has dramatically altered the factual assumptions underlying those decisions.

AI systems can now produce photorealistic images and videos of children who do not exist. This makes it increasingly difficult for investigators to determine whether an image documents the abuse of an actual child or was generated artificially. That uncertainty has real consequences. Law enforcement must devote time and resources to determining whether a depicted child exists. Meanwhile, the growing volume of synthetic imagery can make it harder to identify actual victims who may need immediate intervention.

NCMEC received approximately 67,000 reports involving AI-generated CSAM in 2024—an increase of more than 1,300 percent from the previous year. By the end of 2025, that number had reached 400,000. 

AI also complicates the distinction between “real” and “synthetic” in less obvious ways. Offenders can incorporate the face or body of an identifiable child into synthetic material, create composite images from multiple sources, or alter authentic abuse imagery to make it appear computer-generated. The Seventh Circuit’s concurrence recognized that last danger by citing Enough Abuse’s research on state AI-CSAM laws, which notes that perpetrators may disguise images of real children being abused as computer-generated content.

And there is a deeper question that the record in the AI-CSAM ruling in the Anderegg case could not answer: How “synthetic” is an AI-generated image if the model used to produce it was trained on images of actual child sexual abuse? Researchers have identified known CSAM within datasets used to train image generators. The concurrence recognized that if production of synthetic imagery depends upon, or fuels demand for, material created through the abuse of actual children, the constitutional analysis could look very different. That may become one of the most important issues in the next generation of AI-CSAM litigation.

What Does the AI-CSAM Ruling Mean for State AI-CSAM Laws?

Anderegg does not automatically invalidate state laws addressing AI-generated CSAM.

The Seventh Circuit directly binds federal courts only in Illinois, Indiana, and Wisconsin, and state statutes vary considerably. Some focus on sexual deepfakes involving identifiable real children; others reach wholly synthetic depictions; and some regulate production or distribution more broadly than possession.

At the same time, the constitutional principles underlying Anderegg come from the U.S. Supreme Court and therefore matter nationwide. Challenges are most likely to arise where a law criminalizes nothing more than private possession of wholly synthetic material depicting no actual or identifiable child.

Our research at ENOUGH ABUSE shows that, as of September 2026, 47 states criminalize AI-generated or computer-edited CSAM, while Massachusetts, Ohio, Vermont, and Washington, D.C. still lack comprehensive AI-CSAM provisions in their CSAM statutes. For those states, and for any state revisiting its law, the decision underscores the importance of careful drafting, addressed below.

Could Anderegg (the AI-CSAM Ruling) Reach the Supreme Court?

Two members of the Seventh Circuit panel effectively invited the Supreme Court to revisit the issue. Their concurrence emphasized that AI-generated imagery can now be “virtually indistinguishable” from images of actual abuse and pointed back to Justice Clarence Thomas’s separate opinion in Free Speech Coalition. Thomas predicted in 2002 that technology might eventually become so sophisticated that prosecutors could no longer reliably determine whether an image depicted an actual child. If that happened, he suggested, the government might have a substantially stronger justification for narrowly regulating virtual imagery in order to protect real children. Nearly a quarter-century later, the concurrence suggested that, according to some experts, that moment may have arrived.

Congress is also considering the problem. The Senate unanimously passed the bipartisan  ENFORCE Act  in December 2025, legislation intended to strengthen federal enforcement and align penalties for certain AI-generated CSAM and obscenity offenses with other federal child-exploitation crimes.

Whether Anderegg itself reaches the Supreme Court remains uncertain, but the constitutional questions it raises are unlikely to disappear.

What Parents, Educators, and Policymakers Should Know about the AI-CSAM Ruling

For parents, the most immediate question is simple: Does this decision make it legal for someone to turn my child’s photograph into a sexual image? As explained above, the answer is unequivocally no.

The more important lesson is that ordinary photographs of children can now be copied, manipulated, and repurposed with extraordinary ease. Families do not need to disappear from the internet, but they should think carefully about what they make publicly accessible—especially revealing or highly personal images, and any image paired with a child’s full name, school, location, team, or other identifying information. Some practical precautions include:

• Set family accounts to private, turn off location tagging, and review who can see and share older posts.
• Talk with older children about “nudify” apps, sexual deepfakes, and sextortion before an incident occurs. Make clear that they can come to a trusted adult without fear of punishment.
• If a fake sexual image of your child appears online, preserve the evidence without redistributing the image, report it to the NCMEC CyberTipline, use NCMEC’s free Take It Down tool where applicable, and contact law enforcement.

More recommendations for families can be found in our blog post, “Is it Safe to Post Pictures of Children Online?” 

Schools and youth-serving organizations should plan before an incident occurs. Reports involving AI-generated sexual imagery of students should prompt an immediate, trauma-informed response: follow the same evidence and reporting steps, comply with applicable reporting obligations, support the targeted child, and ensure that technology, harassment, and disciplinary policies expressly address AI-generated sexual imagery. A “fake” image can inflict very real harm.  

Enough Abuse’s Guides for Parents and Online Child Sexual Abuse resources provide additional guidance. In addition, we are providing free, live virtual prevention training, “Enough! Preventing Online Child Sexual Abuse and Exploitation,” open to all. Register here. 

For policymakers, the AI CSAM ruling in the Anderegg case carries a different lesson: laws must keep pace with technology, but they must do so with constitutional precision.

Statutes should distinguish carefully between wholly synthetic material and imagery involving actual or identifiable children and should address conduct closely connected to demonstrable harm, including production, distribution, grooming, sextortion, sexual deepfakes of real children, and the manipulation of authentic abuse imagery. Lawmakers should not assume that courts can solve every problem created by rapidly changing technology after the fact.

The Law Must Catch Up With Technology

The AI-CSAM ruling in the Anderegg case exposes an increasingly difficult conflict between decades-old First Amendment precedent and technology capable of producing imagery earlier courts could scarcely have imagined. The challenge now is to preserve constitutional protections without allowing technological change to undermine society’s ability to identify victims, hold offenders accountable, and protect children from sexual exploitation.

Enough Abuse is working to ensure that state and federal laws keep pace with emerging threats to children. Explore our national map of state laws criminalizing AI-generated and computer-edited CSAM and learn more about sexual deepfakes, AI-generated CSAM, and the laws designed to address them.

If your state has not yet updated its CSAM laws—or if its existing law leaves significant gaps—contact your lawmakers and urge them to act. Protecting children should not depend on whether an abusive image was created with a camera, editing software, or an AI prompt.

Get Help and Report AI-CSAM

Report exploitation → CyberTipline:  report.cybertip.orgor 1-800-843-5678

Remove explicit images of a minor → Take it Down: takeitdown.ncmec.org

Learn More in Other ENOUGH ABUSE Blogs

“Sexual Deepfakes: The Dangers of AI-generated CSAM” 

“Is it Safe to Post Pictures of Children Online?”

“What is Financial Sextortion?”

“Violent Online Networks: A Growing Threat Against Children”

“Roblox: A Playground for Predators” 

 “How to Protect Kids on Social Media” 

About the Author

Jessica Schidlow, Esq. is a children’s rights law and policy expert with extensive experience in statutory reform, appellate advocacy, and systems-level legal analysis. In her previous roles working nationally and internationally, Jessica concentrated on issues related to online exploitation, technology-facilitated abuse, and the intersection of constitutional principles with the rights of children and victims.

About ENOUGH ABUSE

ENOUGH ABUSE® is the nation’s oldest citizen-based child advocacy organization, with a 65-year history of improving the lives of vulnerable children. We work in states, nationally, and with international partners to ensure every child’s right to a childhood free from abuse and exploitation.

For over two decades, our primary focus has been to prevent the sexual abuse of children in their homes, communities and online through our effective prevention trainings, education and advocacy.