A federal appeals court in the United States has ruled that the private possession of AI-generated child sexual abuse material (CSAM) is protected by constitutional free speech rights, provided the images do not depict a real child.
The decision, handed down by the U.S. Court of Appeals for the 7th Circuit on 25 August 2026, marks a significant legal milestone in how technology and the First Amendment intersect. It also highlights a growing legislative divide between American legal protections and the increasingly strict regulations recently adopted in the United Kingdom.
The case involved Steven Anderegg, a software engineer from Wisconsin, who used the generative AI tool Stable Diffusion to create hyper-realistic images. While the court dismissed the charge related to the private possession of the material in his home, the ruling does not grant total immunity. Mr Anderegg still faces active charges concerning the production and distribution of the imagery.
The legal precedent
The court’s decision was rooted in a 2002 U.S. Supreme Court precedent, Ashcroft v. Free Speech Coalition. That ruling established that “virtual” child pornography—material that looks like children but does not involve the use of real children in its creation—is protected speech under the First Amendment.
Legal experts noted that the 7th Circuit’s ruling was a “straightforward reading” of this existing law, despite the rapid advancement of generative AI technology which can now produce imagery indistinguishable from real photography. The court maintained that because no real child was harmed during the creation of the AI images, the state’s interest in protecting children did not override the individual’s right to possess the material privately.
Contrast with UK law
The legal landscape in the UK has moved in the opposite direction. While the American court focused on the absence of a real victim during production, British lawmakers have focused on the potential for AI tools to facilitate further abuse.
The Crime and Policing Act 2026, which received Royal Assent on 29 April 2026, specifically criminalises the possession, supply, and adaptation of “CSA image generators.” Under this Act, it is an offence in the UK to possess a tool specifically designed or adapted to create such material, regardless of whether a real child is depicted.
This follows an environment of heightened regulatory scrutiny in Britain. In January 2026, the communications regulator Ofcom launched an investigation into the social media platform X (formerly Twitter) following reports that an AI chatbot on the service was being used to generate and share sexual imagery involving children.
Practical implications
For UK residents and technology users, the U.S. ruling serves as a reminder that digital content permitted in one jurisdiction can carry heavy criminal penalties in another. The UK’s 2026 legislation was designed to close what ministers described as a “loophole” created by synthetic media, ensuring that the creation and possession of “pseudo-photographs” carry the same legal weight as traditional CSAM.
In the United States, the ruling is expected to be closely watched by civil liberties groups and child protection advocates alike. While the possession charge was dismissed in this instance, the U.S. Department of Justice continues to prosecute the distribution and manufacturing of such material, arguing that even virtual images can be used to groom victims or desensitise offenders.
