Federal Court Ruling on AI-Generated Child Pornography
A federal appeals court has stated that it had to protect the in-home possession of AI-generated child pornography due to a 2002 Supreme Court ruling, even amidst concerns that advancing technology has blurred vital legal lines established nearly 25 years ago.
The Washington Post reported that the Seventh Circuit upheld a lower court’s dismissal of a possession charge against Steven Anderegg, a Wisconsin man alleged to have had hundreds of AI-generated sexually explicit images of children on devices taken from his home.
Anderegg was facing four charges: producing and distributing the AI-generated child sexual abuse material, transferring it to a minor under 16, and possessing it. A federal district court had dismissed only the possession charge, citing First Amendment protections, while the remaining three charges were kept. The Justice Department appealed this dismissal, insisting that the government should be able to prohibit possession of obscene virtual child sexual abuse material, even when no real child was involved in its creation. The Seventh Circuit, however, sided with the lower court’s decision.
This ruling highlights a growing disconnect between Supreme Court decisions and the rapid evolution of AI technology. The appeals court acknowledged its obligation to follow the 2002 Supreme Court decision in Ashcroft v. Free Speech Coalition, which had rejected limits on sexually explicit representations of fictional children when no actual minors were involved. Nonetheless, two judges cautioned that today’s AI can create images that closely resemble those depicting the abuse of real children.
Judge John Lee pointed out that in the earlier ruling, the Supreme Court considered First Amendment protections for virtual child sexual abuse material, but technology has advanced dramatically since then, perhaps beyond what could have been imagined back then. “We do have concerns about the lines these cases draw, but we are not free to redraw them ourselves,” he remarked.
Generally, the government cannot criminalize private possession of obscene material in someone’s home but can bar possession of child sexual abuse material involving actual minors since such materials exploit real victims. The Supreme Court’s 2002 decision complicated matters when it ruled against restrictions on sexually explicit depictions of fictional children created without using real minors, reasoning that the need to protect children from exploitation doesn’t apply when no real child is involved.
Hans von Spakovsky, a senior fellow at Advancing American Freedom, noted that the appeals court’s decision was a necessary outcome under the First Amendment due to the precedent set by the Supreme Court. He echoed the concerns expressed by Judges Lee and Kolar, suggesting that the rapid advancement of AI technology might necessitate a reevaluation of prior legal standards.
He pointed out, “AI-generated images are now nearly indistinguishable from the actual thing, and some of them could be produced based on real child sexual abuse material.” He emphasized that technology might now outpace the reasoning behind earlier judicial decisions.
The Supreme Court previously ruled that altering images of real children into sexually explicit poses would not be covered by the protection for possession of child sexual abuse material involving fictional children since this practice impacts the interests of real minors. Justice Department lawyers have cautioned that the proliferation of AI-generated child sexual abuse material could complicate efforts to identify genuine victims, as more realistic artificial imagery becomes increasingly hard to differentiate from actual abuse depictions.
In their joint statement, Lee and Kolar expressed additional concerns based on research indicating that numerous child sexual abuse material images were present in a major dataset used to train image-generating AI. They suggested this could link AI-generated content to the exploitation of real children, although they emphasized that this particular issue was not argued in the current case and the evidence presented was insufficient for a resolution.
Justice Clarence Thomas had anticipated such dilemmas over two decades ago, warning in the 2002 ruling that technological progress might eventually lead to virtual depictions becoming indistinguishable from actual child sexual abuse material involving real victims.
Thomas wrote, “If technological advances thwart prosecution of ‘unlawful speech,’ the Government may well have a compelling interest in barring or regulating a narrow category of ‘lawful speech’ to enforce laws against material produced through the abuse of real children.”
In related news, a victim who was exploited as a child to create child pornography has filed a lawsuit against Elon Musk’s xAI, claiming that her images were utilized to train its Grok AI. She receives notifications from the U.S. Department of Justice Victim Notification System whenever her abuse material appears in new locations. The Canadian Centre for Child Protection alerted her that AI-generated child sexual abuse material featuring her has been identified within xAI’s systems. According to the lawsuit, discussions about generating AI-based child sexual abuse material featuring her and other known victims occurred on various online forums.
The lawsuit alleges that xAI stores and reuses images generated by Grok to further train the AI model. Doe’s attorneys described the material as being “that same material,” referring to the child sexual abuse material that investigators say contributed to Grok’s outputs. The complaint states, “CSAM depicting Plaintiff with its longstanding well-known hash values has been used as a part of the dataset used by xAI.”
While AI has significant potential benefits, it also poses serious risks, including the emergence of AI-generated child pornography. In addressing this topic, Breitbart News social media director Wynton Hall has produced a bestseller titled Code Red: The Left, the Right, China, and the Race to Control AI, aiming to provide insights on how the MAGA movement could approach AI in a way that promotes humanity without ceding control to Silicon Valley or China.
