Attorneys general from all 50 states and four territories are asking Congress to address a growing concern: generative AI can be used to create child sexual abuse material (CSAM), and existing laws may not clearly cover every AI-generated image. Their proposal pairs an ongoing expert review with possible updates to the law.
Why AI-generated images concern prosecutors
Some open-source image-generation tools can be run locally, and the attorneys general say that can leave them without safeguards against creating sexualized images of children. A user can generate an image by entering a short description, while add-ons and other tools can extend what these systems do.
The letter also raises the risk of deepfakes involving identifiable children. AI tools could use photographs of abused children to generate new sexual images, or combine images of previously unharmed children found online with images of abuse. In either case, a child’s likeness could be drawn into material they never consented to create.
The concern is not limited to whether an image depicts a real act of abuse. The attorneys general argue that even unrealistic AI-generated material could normalize child abuse and support the market for child exploitation. That view puts the focus on potential effects on children and families as well as on how the images were made.
The legal question is difficult
Congress faces a challenge in drawing rules that protect children without unintentionally restricting people’s ability to use AI to generate legal images. Technical limits could affect both kinds of use, and widely used models such as Stable Diffusion have already been released. The letter acknowledges that governments cannot simply reverse that release.
There is also a complicated legal backdrop. The article notes that some fictional depictions of CSAM are already illegal in the United States, while the attorneys general call for existing restrictions to explicitly cover AI-generated CSAM. How those rules would apply to different kinds of generated material is part of the issue they want lawmakers to examine.
The wider debate involves free speech and privacy as well as child safety. Past technology measures intended to combat CSAM have had controversial side effects, including potential effects on the privacy and rights of law-abiding people. The broad appeal of protecting children can also be used to support censorship, making careful consideration of any new restrictions important.
What the attorneys general propose
Rather than asking Congress to start with a specific technical fix, the attorneys general recommend an expert commission focused on AI and child exploitation. They say it should operate on an ongoing basis because the technology is changing quickly, and should identify ways AI could be used to exploit children and propose ways to prevent and address that harm.
After considering the commission’s recommendations, Congress should consider expanding existing CSAM restrictions to explicitly include AI-generated material. The attorneys general say that would give prosecutors clearer tools to pursue child exploitation cases.
Their proposed sequence makes the commission a way to build a current understanding of the problem before lawmakers settle on a response. It also recognizes that AI policy designed for the technology as it exists today may need to be revisited as the tools change.
A call for action, with open questions
The request reflects agreement among a politically diverse group of attorneys general that AI-generated child sexual abuse material deserves attention. Their letter describes the issue as urgent, while its recommendations leave the details of legislation for Congress to determine.
Any response will have to weigh the potential harm of generated images against the risks of rules that reach lawful AI use, privacy, or speech. The attorneys general are asking lawmakers to study those questions and make sure child protection remains central as generative AI develops.