AI image generators are making it possible to produce lifelike images depicting child sexual abuse, adding a new challenge for investigators and child safety organizations. They must contend with the growing volume of fabricated material while continuing to identify real victims and investigate actual abuse.
More images complicate the search for victims
The Washington Post reported an increase in AI-generated abuse images, citing research by Thorn, a nonprofit child safety group. Investigators fear that the new material could overwhelm systems designed to track abuse imagery and help locate victims.
Each image may require additional work to determine whether it depicts a real child or was generated by AI. That classification task sits alongside investigators’ existing work. If fabricated images flood tracking systems, they could make it harder to find the material that points to real victims.
Yiota Souras, director of legal affairs at the National Center for Missing and Exploited Children, confirmed a sharp increase in AI-generated abuse images. She also warned that mass-produced images could be used to encourage real children to imitate what they see.
Online forums share tools and instructions
The report describes an arms race around generative AI in pedophile forums. Thousands of images have been found on Dark Web forums, where users also share instructions for making their own images with open source models.
Avi Jager, director of child protection and human exploitation at ActiveFence, said a survey of one forum with more than 3,000 participants found that about 80 percent had used or planned to use an AI image generator for this purpose. ActiveFence helps social media and streaming sites prevent child pornography on their platforms.
The reported activity points to a challenge that extends beyond the creation of images. Instructions and tools shared among forum participants can help spread the practice, while platforms and investigators need to identify and respond to the resulting material.
Legal treatment remains uncertain
Child advocates and U.S. justice officials say AI-generated images should be punishable by law. But the source reports that no rulings yet establish the level of punishment.
One difficulty is that fabricated images do not include real children. That leaves an open question about whether they violate federal child-protection laws. The uncertainty complicates efforts to determine how such cases should be handled, even as advocates call for legal consequences.
Detection and prevention may help
One possible response is to develop systems that can reliably distinguish AI-generated images from real photographs. If those systems can be integrated into online recognition tools, they could help investigators sort material and focus attention on images that may involve actual victims.
Another approach is to label images inside the software that creates them. Clear labels could help identify generated material as it moves online, though the source presents this as a potential measure rather than an existing solution.
Separately, the FBI warned in early June about the use of generative AI to make explicit images from existing pictures, including images taken from social media. The agency has seen an increase in these so-called "sextortion victims" since April 2023.
Together, these concerns show why image generation creates demands for investigators, online services and lawmakers. They need to identify fabricated images without losing sight of real children who may be experiencing abuse or exploitation.