Paedophiles are leveraging AI to generate depictions of celebrities when they were young.
The IWF reported that predators are distributing visual depictions of a widely known female singer appearing as a child.
The charity revealed that images of child actors are being changed for illicit purposes on a particular dark web forum.
Using bespoke image generators, actual victims of child sexual abuse are having hundreds of images of them created.
The IWF's latest report, which is aimed at increasing knowledge about the risks posed by paedophiles who make use of AI systems that produce images on the basis of basic text instructions, provides the information.
Since the arrival of these sophisticated image creation programs in the public domain, researchers have cautioned that there is a possibility of them being abused to make illegal images.
In May, Home Secretary Suella Braverman and US Homeland Security Secretary Alejandro Mayorkas released a joint declaration affirming their intent to address the "dreadful increase in Artificial Intelligence generated pictures of minors that are being abused by paedophiles".
The IWF's report explains that for a duration of one month, researchers monitored AI imagery on a single child abuse website on the darknet and identified close to 3,000 synthetic pictures that break UK law.
Analysts have noted a recent development where predators take a single photograph of a widely-recognized victim of child abuse and then generate many more images of the same person in different situations of sexual abuse.
While examining a folder, investigators discovered 501 pictures of an actual minor who was roughly nine or ten years old when subjected to sexual mistreatment. This folder had not only the images, but also an AI model that had been perfected in order to generate replicas of the victim’s pictures.
According to the IWF, certain images, including those featuring celebrities when they were still young, are extremely credible and would be difficult to differentiate from the genuine article by someone who is inexperienced.
Analysts viewed pictures of mainly female singers and movie stars that had been digitally-altered with imaging software to give them the appearance of being young children.
It was not revealed in the report which celebrities were targeted.
The charity declared that they were broadcasting the research to ensure the problem was included in discussion at the upcoming UK government AI Summit, to be held at Bletchley Park next week.
Within a period of one month, the IWF (Internet Watch Foundation) examined 11,108 AI-generated images that had been posted on a darknet online forum devoted to the sharing of child abuse materials.
In June, the IWF expressed concern that predators were beginning to investigate the use of AI to create depraved images of children, and now they have confirmed that these fears have become a reality.
Susie Hargreaves OBE, the head of the IWF, expressed that their greatest fears had been realized.
Back in the beginning of the year, we predicted that AI-generated images would soon be difficult to distinguish from genuine pictures of children who had been victims of sexual abuse, and that larger numbers of this type of image would become prevalent. This has now become a reality.
The IWF reiterates the fact that AI images bring about real harm -- the consequences of which are not felt by children directly. However, such images normalize the predatory behavior of adults, in addition to using up the resources of the police who have to probe children who do not really exist.
Investigators are investigating new types of crimes, which raises fresh challenges for police departments.
As one example, the IWF located multiple photos of two young women who had been photographed for a non-nude modelling session, but those images had been changed to include Category A sexual abuse content.
The truth of the matter is that they have now become the victims of Category A offences which never occurred.
Comments