Criminals create new sexual images of children through AI – Utilizing previous material of victims

Sexual predators operating on the dark web are using artificial intelligence (AI) to create sexual images of children, leveraging previously leaked material.

Child protection groups have identified conversations between perpetrators about creating new images based on older child sexual abuse material (CSAM). These perpetrators are obsessed with childvictims who are referred to as “stars” in predator communities due to the popularity of their images.

“The communities of people who market this material are excited about individual children,” said Sarah Gardner, CEO of the Heat Initiative, a nonprofit focused on child protection. “They want more content from these children, which is what AI has now enabled them to do.”

These children may now be adults, but because of artificial intelligence, new images of them are being created and more people may now be seeing sexual content depicting them as children. They fear that images of them circulating online or in criminal communities could threaten their lives and careers.

See also: Twitch allows filtering and blurring of sexual content

Megan, a CSAM victim, says the possibility of AI being used to create images of her has become an increasingly stressful prospect over the past 12 months, even though her own abuse occurred a decade ago.

"AI gives perpetrators the opportunity to create even more situations of my abuse to feed their own fantasies and their own versions," he said. "The way they could manipulate my images with AI could give the false impression that it was not harmful or that I enjoyed the abuse."

Since this material is mostly circulated on the dark web, there is an anonymity and child safety groups have few means to request the removal of these images or report users to law enforcement.

The groups have called for legislation that goes beyond criminalization to prevent AI and other forms of child sexual abuse from being produced. However, they are pessimistic and fear that little can be done to enforce bans on the creation of new child sexual abuse images. Furthermore, encrypted messaging, which are now often the default options, allow sexual predators to exchange content.

Creating new CSAM and reviving old sexual content with AI

The Guardian has seen several excerpts from these dark conversations. Forum members are encouraged to create new sexual images with AI to share with others. Many said they were excited by the prospect of new material, while others said they were not interested because it did not depict actual child abuse.

See also: Taylor Swift: Inappropriate AI-generated photos spread on social media

A message from November 2023 read: “Could you use AI to recreate the beautiful images of former CP stars [child porn] [victim name] and [victim name] and put them in certain scenes – like [victim name redacted] in a traditional Catholic schoolgirl uniform in Elementary School and [victim name redacted] in a cheerleader uniform in High School?”

In another chat room conversation, sexual predators also discussed using artificial intelligence to digitally reconstruct popular child exploitation materialthat was decades old and of low quality.

“Wow, you’re awesome,” one member wrote to another in January. “I appreciate your effort, keep upgrading the classic videos.”

In the past, photo editing software has been used for similar purposes, but AI gives criminals new opportunities to create more realistic sexual (abuse) images of children.

Much of this activity focuses on the so-called "stars".

“In the same way that there are celebrities in Hollywood, in these online communities, there is a celebrity-like ranking of some of the favorite victims,” said Jacques Marcoux, director of research and analysis at the Canadian Centre for Child Protection. “These groups of offenders know everyone and they list them.”

"The perpetrators eventually exhaust all the footage of a particular victim," Marcoux said. "So they can get an image of a victim that they like, and they can make that victim do different poses or do different things."

The data confirms the phenomenon of criminals engaging with “stars.” In a 2020 assessment by the National Center for Missing and Exploited Children, Meta reported that just six videos accounted for half of all child sexual abuse material shared and reposted on Facebook and Instagram.

See also: Deepfake & Sextortion: Fraudsters create inappropriate content using your social media

Real-life Hollywood celebrities are also potential targets for victimization. Conversations were found that spoke of the potential creation of inappropriate child material by celebrities, including teen idols from the 1990s who are now adults.

How does AI-generated child sexual abuse material spread?

The use of artificial intelligence by sexual predators became widespread in late 2022. That same year, the LAION-5B database was released , which contains more than 5 billion images and can be used by anyone to train artificial intelligence models.

However, a report by Stanford University revealed that hundreds of child sexual images had been included in the LAION-5B database and are now being used to train popular AI image-creation models to create CSAM.

"My body will never be mine again and that's something that many survivors have to deal with," said people who were sexually abused as children and now the images are being recreated through AI.

Experts say they have seen a shift towards encrypted private messaging services like WhatsApp, Signal and Telegram for the dissemination and access of CSAM. However, much of CSAM is still shared outside of mainstream channels on the dark web. In an October 2023 report, the Internet Watch Foundation (IWF) says it found more than 20,000 AI-generated child sexual images posted on a single forum.

Some sexual predators mistakenly believe that displaying AI-generated CSAM may be more ethical than “real-life” material, experts said.

«One of our concerns is their ability to justify their behavior, because these are in some ways images of a victimless crime that doesn't involve real-world harm. Some of the people who call us make an argument to minimize the gravity of what they're doing.».

How will AI-generated sexual images of children be restricted?

In many countries, including the US and UK, decades-old laws already criminalize any CSAM created using AI. Also, the pornographic depictions of Taylor Swift made by AI that were released earlier this year led to legislation in the US to regulate such deepfakes.

See also: Google SafeSearch: Will soon blur inappropriate images by default

Child safety and technology experts have been pessimistic about whether it is possible to prevent the production and distribution of AI-generated CSAM, noting that much of the production goes unnoticed by authorities.

How can parents protect their children from such threats?

• Parents and guardians can protect their children from such threats by educating them about safe internet use. It is important to explain the risks involved and teach them how to recognize suspicious behavior and content.

• Using parental control software can help monitor children's online activity and block access to inappropriate websites. These tools can provide alerts for suspicious activity and help parents maintain control.

• Open and honest communication is critical. Parents need to create an environment where children feel comfortable talking about their online experiences and reporting any concerning contact or content.

• Educating children about protecting their privacy is also important. They need to learn not to share personal information, photos , or videos with strangers and understand the consequences of sharing such material.

• Parents need to be aware of the latest technological developments and new threats that may arise. Staying informed and participating in seminars or support groups can help them stay prepared.

Source: www.theguardian.com