Posted by [By Jaster Francis]
Few things have become more polarizing than artificial intelligence. Some hate it, refusing to use the technology for any reason. Some love it, constantly asking models for help with the smallest tasks. And recently, the Seventh Circuit addressed a third category: those who use AI to commit federal crimes.
In US v. Anderegg, No. 25-1354, 2026 WL 2490623 (7th Cir. Aug. 25, 2026), the Seventh Circuit tackled the case of Steven Anderegg, a software engineer with a professional background in AI models. According to the allegations in the indictment, Anderegg produced, distributed, and possessed visual depictions of minors engaged in sexually explicit conduct and transferred that material to a minor. Those visual depictions were allegedly developed with Stable Diffusion, a “generative artificial intelligence” software that allowed Anderegg to create hyper-realistic visual depictions of prepubescent children. And Instagram messages—which involved an account allegedly linked to Anderegg—appeared to contain computer-generated CSAM that had been sent to a minor’s account.
The government charged Anderegg with violating various statutes, among them 18 U.S.C. § 1466A(b)(1) and (d)(4). That statute criminalizes, among other acts, “knowingly possess[ing] a visual depiction of any kind, including a drawing, cartoon, sculpture, or painting, that . . . depicts a minor engaging in sexually explicit conduct; and . . . is obscene” where any visual depiction involved in the offense has been shipped and transported in interstate and foreign commerce.
Anderegg challenged this charge, arguing that prosecuting him for possessing the AI-generated images would violate his First Amendment rights. The district court agreed, dismissing the possession charge as violative of the First Amendment but letting the other counts proceed. And in an interlocutory appeal, the government urged the Seventh Circuit to reverse the district court’s decision.
But the Seventh Circuit affirmed the dismissal. Invoking decades-long Supreme Court case law, the Seventh Circuit agreed that § 1466A(b)(1) violated the First Amendment as applied to Anderegg. The court highlighted that, unlike child sexual abuse material (“CSAM”) that depicts real children, AI-generated CSAM does not require the victimization of real people. As such, the court reasoned, Supreme Court cases allowing the prosecution of in-home CSAM possession are inapplicable to the type of AI-generated material that Anderegg allegedly possessed.
The Seventh Circuit also decided that AI-generated CSAM was more akin to obscene materials generally, not actual CSAM. And because Anderegg was being prosecuted for possessing obscene materials within the four walls of his home—and the Supreme Court had long rejected prosecutions for in-home, private possession of obscene materials—the Seventh Circuit held that Anderegg’s prosecution for possession of obscene materials was unconstitutional.
This case is an important reminder that as technology evolves, so too should our litigation strategies.
