You may have seen an alarming claim circulating online: “The Supreme Court just ruled that AI child pornography is legal in the United States.”
No, that isn’t what happened. First, the Supreme Court did not issue the ruling a federal appeals court did second, the court did not broadly legalize AI-generated child sexual abuse material (CSAM), either. Instead, it upheld the dismissal of one possession charge against one defendant under the specific facts of one case.
Correcting that misinformation is important, but so is the actual ruling.
In fact, the case exposes a difficult decision courts, lawmakers, technology companies, and child-safety advocates increasingly face when decades-old legal distinctions collide with artificial intelligence capable of creating sexual abuse imagery that can look virtually indistinguishable from images of real children.
Fight the New Drug’s position remains clear: children should never be sexualized or treated as sexual products. Whether technology creates, alters, or distributes the imagery, AI-generated CSAM (child sexual abuse material) raises serious concerns for child safety, exploitation, consent, and the normalization of sexual abuse.
Here’s what the court actually decided—and why the case matters.
The Case Wasn’t a Supreme Court Case
The case is United States v. Anderegg, decided on August 25, 2026, by the U.S. Court of Appeals for the Seventh Circuit.
According to the government’s allegations, Steven Anderegg used a generative AI program to produce hundreds of photorealistic, sexually explicit images depicting prepubescent children. Authorities investigated after Meta reported a transmission of suspected CSAM to the National Center for Missing & Exploited Children’s CyberTipline. The government alleged that an Instagram account linked to Anderegg sent AI-generated imagery to another minor’s account.
Prosecutors charged Anderegg with four federal offenses involving production, distribution, transfer to a minor under 16, and possession. The important part: the district court dismissed only the possession charge. The production, distribution, and transfer charges remained. He did not get off the hook, nor did the court suddenly legalize AI CSAM.
The government simply appealed the dismissal of the possession charge alone; the Seventh Circuit affirmed the dismissal.
So the court did not announce that AI-generated CSAM is broadly legal. It considered whether one particular federal possession statute could constitutionally apply to Anderegg under the circumstances alleged by the government. The government’s concession was also critical: the images underlying the charge did not depict an actual child and could not be linked to one.United States v. Anderegg, No. 25-1354 (7th Cir. Aug. 25, 2026). https://media.ca7.uscourts.gov/cgi-bin/OpinionsWeb/processWebInputExternal.pl?Path=Y2026/D08-25/C:25-1354:J:Lee:autconi:T:fnOp:N:3597567:S:0&Submit=DisplayCopy
Why Did a Court Dismiss the One Possession Charge?
The answer requires going back decades before modern generative AI existed. The Seventh Circuit concluded that three Supreme Court cases shaped the outcome. In Stanley v. Georgia, decided in 1969, the Supreme Court held that the government could not criminalize someone’s mere private possession of obscene material inside their home.Stanley v. Georgia, 394 U.S. 557 (1969). https://supreme.justia.com/cases/federal/us/394/557/Copy
Then came Osborne v. Ohio in 1990.
That case involved CSAM depicting actual children. The Supreme Court allowed the government to prohibit its possession, including possession in the home, because the sexual exploitation of real children changes the constitutional analysis. CSAM involving real children does not merely depict abuse; its production requires the sexual exploitation of children, and continued circulation can perpetuate the harm.Osborne v. Ohio, 495 U.S. 103 (1990). https://supreme.justia.com/cases/federal/us/495/103Copy
Then, in 2002, the Supreme Court considered virtual depictions in Ashcroft v. Free Speech Coalition. Congress had attempted to expand federal prohibitions to include some sexually explicit depictions that appeared to feature minors, even when no actual children participated in their creation. The Supreme Court found portions of that law unconstitutionally overbroad.Ashcroft v. Free Speech Coalition, 535 U.S. 234 (2002). https://supreme.justia.com/cases/federal/us/535/234Copy
That distinction became crucial in Anderegg. The government conceded that the images involved in the charge did not depict identifiable real children. And because the possession allegedly occurred in the home, the Seventh Circuit concluded that Stanley and Free Speech Coalition controlled the constitutional question before it.
In other words, the appeals court didn’t announce a new constitutional right to AI-generated CSAM. It concluded that existing Supreme Court precedent required the result in this specific case.
Even the Judges Said Today’s Technology Complicates Those Old Legal Lines
The Seventh Circuit didn’t ignore how dramatically technology has changed. The court explicitly recognized the problem:
“We now live in an age where GenAI models can render images depicting the abuse of virtual children that are virtually indistinguishable from those depicting the abuse of actual children.”
The judges noted that Free Speech Coalition came down nearly 25 years ago, when today’s image-generation technology was likely unimaginable. And they acknowledged their own discomfort with applying those precedents to modern AI:
“We have some concerns about the lines these cases draw, but we are not free to redraw them ourselves.”
Judges Thomas Lee and Michael Kolar went further in a separate concurrence, writing that courts “would benefit from additional guidance from the Supreme Court” about how the First Amendment applies to virtual CSAM in light of today’s technology. That’s an important part of this story. The appeals court wasn’t necessarily endorsing AI-generated CSAM or concluding that it poses no harm. The judges said lower courts must follow binding Supreme Court precedent unless the Supreme Court changes it.
The legal question now exists in a technological world that looks radically different from the one in which that precedent developed.
“AI-Generated” Doesn’t Necessarily Mean No Real Child Was Involved
One reason that distinction has become so complicated is that modern AI imagery doesn’t fit neatly into categories of “real” and “fake.”
AI can transform an ordinary photograph of a real child into sexually explicit imagery. It can generate new images resembling known survivors of child sexual abuse. Models can also learn from enormous collections of existing images.Researchers have already documented CSAM within datasets used to train generative AI.
In 2023, researchers at the Stanford Internet Observatory investigated the LAION-5B dataset, which researchers and developers had used to train popular image-generation models. Their investigation identified hundreds of validated instances of CSAM within the dataset and thousands of suspected instances that warranted further investigation.Thiel, D., Stroebel, M., & Portnoff, R. (2023). Identifying and eliminating CSAM in generative ML training data and models. Stanford Internet Observatory. https://purl.stanford.edu/kh752sm9123Copy
That doesn’t mean every AI-generated image relies directly on CSAM, or that every generative AI model contains abusive material in its training data. It does mean the phrase “completely artificial” can obscure a complicated technological supply chain.
The Seventh Circuit itself highlighted this issue in its concurrence, noting research showing that a major AI training dataset contained hundreds of CSAM images. The judges observed that this possibility could complicate the assumption that virtual material exists entirely separately from the abuse of actual children.
AI Is Making It Trickier to Know What’s Real
The legal distinction between a real child and a virtual child becomes harder to enforce when the human eye may struggle to tell the difference. The Internet Watch Foundation’s 2026 report on AI-generated CSAM documents rapidly increasing realism and scale.
In 2025, IWF analysts identified 8,029 realistic AI-generated images and videos they assessed as criminal under UK law. The organization identified 3,443 AI-generated abuse videos that year—more than 260 times the number identified in 2024. Some of this imagery incorporates or depicts recognizable real victims, and increasingly realistic synthetic material can create another problem: investigators must determine whether an apparent victim actually exists.
The Seventh Circuit noted concerns that an influx of AI-generated material can consume law-enforcement resources and complicate efforts to identify children who are currently experiencing abuse. That means the harms of AI-generated CSAM cannot always be evaluated by asking only whether a child physically participated in creating a particular image.
Research Warns Against Treating Fully Synthetic CSAM as Harmless
What if investigators could conclusively establish that an image was entirely synthetic—that no identifiable child appeared in it and no particular abusive image directly produced it? Researchers still warn against describing that material as harmless.
A 2026 review published in AI & Society examined emerging evidence around AI-generated CSAM. The authors reviewed psychological and criminological research alongside technical literature, civil-society reports, and law-enforcement evidence.
They identified several possible harms.
AI-generated CSAM can facilitate grooming, coercion, and sexual extortion. It can depict known survivors and revictimize them. Researchers also raised concerns that it could lower barriers to engaging with CSAM, contribute to desensitization, reinforce distorted beliefs about children and sexual abuse, and provide pathways toward increasingly extreme material.
Importantly, the authors describe their work as a narrative review, rather than a systematic review, because AI-generated CSAM remains an emerging phenomenon and the available evidence comes from varied sources. That limitation matters: researchers still need more empirical evidence to determine how these risks operate and for whom. But the absence of complete evidence doesn’t make the documented risks disappear.
As the researchers explain, focusing only on whether a child experienced abuse at the exact moment an AI image was generated creates a misleading comparison. AI-generated material can exist within broader systems of exploitation even when a particular image is synthetic.Ó Ciardha, C., Buckley, J., & Portnoff, R. S. (2026). AI-generated child sexual abuse material: What’s the harm? AI & Society. https://doi.org/10.1007/s00146-026-02932-yCopy
AI-Generated Sexual Content Is Already Being Used Against Real Children
The concern isn’t hypothetical. Someone can take an ordinary photograph of a real child—something posted to social media, shared with classmates, or taken at school—and use AI to fabricate sexual imagery. That material can then become a weapon for harassment, humiliation, grooming, coercion, or sexual extortion.
Research into the broader AI “nudify” ecosystem shows how accessible these tools have become. A 2026 investigation by the Institute for Strategic Dialogue identified 181 websites that allowed users to upload photographs of real people and generate synthetic nude or sexually explicit images or videos. Researchers estimated that these sites collectively attracted more than 40 million unique visitors per month during the study period.
The imagery may be synthetic, but when someone manipulates the likeness of a real child without consent, the victimization isn’t imaginary. A child can still experience humiliation, fear, harassment, reputational harm, coercion, or exploitation because of an image depicting something that never physically happened.
That’s one reason conversations about AI-generated sexual abuse need to move beyond the simplistic question, “But is the picture real?”
The person being targeted may be very real.
Could AI-Generated CSAM Normalize the Sexualization of Children?
Researchers also worry about what happens when technology makes sexualized representations of children easier to generate, customize, and consume. The 2026 AI & Society review warns that synthetic CSAM may contribute to normalization and desensitization while reinforcing cognitive distortions related to child sexual exploitation. That does not establish that every person who encounters AI-generated CSAM will commit contact abuse.
Researchers are examining a more nuanced concern: whether easily accessible synthetic material could lower barriers for some users, reinforce sexual interest in children, facilitate movement toward more extreme material, or undermine factors that otherwise discourage offending.
We don’t need to claim that one image inevitably causes another crime to recognize the danger of building technology ecosystems that make the sexualization of children easier, faster, more realistic, and more customizable.
The Technology Changed. The Need to Protect Children Didn’t.
So, did the Supreme Court legalize AI-generated CSAM?
No.
The Supreme Court didn’t decide United States v. Anderegg at all.
The Seventh Circuit upheld the dismissal of one possession charge because it concluded that Supreme Court precedents involving private possession and virtual imagery controlled the specific constitutional question before it. Other charges against Anderegg—including production, distribution, and transfer of the material to a minor—were not dismissed by that ruling.
The case exposes how quickly generative AI has complicated legal categories created in another technological era. Courts must now confront imagery that can be virtually indistinguishable from depictions of real abuse, technology that can sexualize photographs of actual children, models that may learn from datasets contaminated with abuse imagery, and synthetic content that can burden efforts to identify real victims.
The law will continue grappling with where constitutional boundaries belong. Our standard for human dignity doesn’t need to be so complicated.
Children are people, not sexual products. Their likenesses, bodies, vulnerability, and identities should never become material for someone else’s sexual gratification.
As AI evolves, protecting children will require thoughtful legislation, responsible technology design, effective platform safeguards, better digital literacy, continued research, and honest conversations that resist both misinformation and minimization.
We can accurately explain what this court decided while still recognizing what the decision reveals: our technology has changed dramatically, and our efforts to protect children must keep pace.
Your Support Matters Now More Than Ever
Most kids today are exposed to porn by the age of 12. By the time they’re teenagers, 75% of boys and 70% of girls have already viewed itRobb, M.B., & Mann, S. (2023). Teens and pornography. San Francisco, CA: Common Sense.Copy —often before they’ve had a single healthy conversation about it.
Even more concerning: over half of boys and nearly 40% of girls believe porn is a realistic depiction of sexMartellozzo, E., Monaghan, A., Adler, J. R., Davidson, J., Leyva, R., & Horvath, M. A. H. (2016). “I wasn’t sure it was normal to watch it”: A quantitative and qualitative examination of the impact of online pornography on the values, attitudes, beliefs and behaviours of children and young people. Middlesex University, NSPCC, & Office of the Children’s Commissioner.Copy . And among teens who have seen porn, more than 79% of teens use it to learn how to have sexRobb, M.B., & Mann, S. (2023). Teens and pornography. San Francisco, CA: Common Sense.Copy . That means millions of young people are getting sex ed from violent, degrading content, which becomes their baseline understanding of intimacy. Out of the most popular porn, 33%-88% of videos contain physical aggression and nonconsensual violence-related themesFritz, N., Malic, V., Paul, B., & Zhou, Y. (2020). A descriptive analysis of the types, targets, and relative frequency of aggression in mainstream pornography. Archives of Sexual Behavior, 49(8), 3041-3053. doi:10.1007/s10508-020-01773-0Copy Bridges et al., 2010, “Aggression and Sexual Behavior in Best-Selling Pornography Videos: A Content Analysis,” Violence Against Women.Copy .
From increasing rates of loneliness, depression, and self-doubt, to distorted views of sex, reduced relationship satisfaction, and riskier sexual behavior among teens, porn is impacting individuals, relationships, and society worldwideFight the New Drug. (2024, May). Get the Facts (Series of web articles). Fight the New Drug.Copy .
This is why Fight the New Drug exists—but we can’t do it without you.
Your donation directly fuels the creation of new educational resources, including our awareness-raising videos, podcasts, research-driven articles, engaging school presentations, and digital tools that reach youth where they are: online and in school. It equips individuals, parents, educators, and youth with trustworthy resources to start the conversation.
Will you join us? We’re grateful for whatever you can give—but a recurring donation makes the biggest difference. Every dollar directly supports our vital work, and every individual we reach decreases sexual exploitation. Let’s fight for real love:
