Court Ruling Exposes Gap in AI-Generated CSAM Prosecution

Federal court blocks AI-generated CSAM prosecution

The Seventh Circuit Court of Appeals last week upheld the dismissal of federal charges against a defendant accused of possessing obscene AI-generated child sexual abuse imagery, marking the government's first significant appellate setback in prosecuting AI-created CSAM.

Steven Anderegg allegedly used Stable Diffusion to produce hyperrealistic images of prepubescent children in explicit sexual acts and distributed them via Instagram direct message to a minor's account. While prosecutors charged him with producing, distributing, and possessing CSAM under federal law, the trial court dismissed the possession charge—a decision the appeals court reluctantly affirmed.

The ruling stems from a collision of Supreme Court precedents spanning nearly five decades. While the Court has long held that obscenity lacks First Amendment protection and created exceptions allowing prosecution of actual child sexual abuse imagery regardless of obscenity standards, the 2002 Ashcroft v. Free Speech Coalition decision limited those exceptions to imagery involving real children. That means virtual CSAM that doesn't meet local obscenity standards receives constitutional protection, and even obscene virtual CSAM cannot be prosecuted for mere possession under Stanley v. Georgia's privacy rationale.

Why it matters

The explosion of AI-generated CSAM directly undermines law enforcement's ability to combat child exploitation. Reports to the National Center for Missing and Exploited Children surged from 4,700 AI-generated CSAM images in 2024 to over 400,000 in 2025. This flood of synthetic imagery makes it harder to detect actual abuse documentation and creates a ready-made defense for perpetrators who can claim real images are AI-generated. One study found users prompted Grok to create over 3 million sexualized images in just eleven days, including more than 23,000 constituting CSAM.

Technology has outpaced precedent

When the Supreme Court decided Ashcroft in 2002, virtual images were difficult to produce and visibly inferior to photographs. Justice Clarence Thomas found arguments about defendants using a "computer-generated" defense persuasive but too speculative at the time. Today's AI tools generate indistinguishable imagery quickly and inexpensively, making that defense inevitable.

Judges Milton C. Lee and Joshua P. Kolar wrote in their concurrence that "given the significant advances in computer-generated image technology, particularly in the field of artificial intelligence, we would benefit from additional guidance from the Supreme Court regarding the intersection between the First Amendment and virtual CSAM."

The Supreme Court has shown willingness to revisit early-internet-era decisions. Last year it upheld a Texas age verification law for online pornography despite effectively overturning similar federal laws from 1997 and 2004, acknowledging that internet technology "has continued to evolve at a rapid pace." The Court noted capabilities that "would have been unimaginable" when those earlier cases were decided—language that applies equally to AI-generated CSAM at scale.

These details were first reported by the American Enterprise Institute.