AI-Generated Child Sexual Abuse Images 'Threatens To Overwhelm The Internet

AI can do a lot of things.

Thanks to the abundance of training materials, paired with hardware that is increasingly becoming more powerful, and software that is tweaked and optimized, AI is only going to get better and smarter as time passes.

The thing is, not everything AIs do is in line with the society.

In this case, a UK-based internet watchdog, the Internet Watch Foundation (IWF), said that it has found around 3,000 AI-generated child sexual abuse images (CSAM) that broke the laws in the UK.

These images are threatening the internet, it said.

This happens because of a few reasons.

First, they come in large numbers. Using generative AI tools, it's very easy for practically anyone to create some sort of CSAM.

Second, some specialized tools can be used to "de-age" celebrities to make them look like children, and then use the aforementioned tools to depict them as CSAM.

According to Dan Sexton, the chief technology officer at the IWF, people were using Stable Diffusion, a publicly available image-generating tool to help them create CSAM.

"We have seen discussions around the creation of content using Stable Diffusion, which is openly available software," Sexton said.

Stability AI, the UK company that created Stable Diffusion, has said it "prohibits any misuse for illegal or immoral purposes across our platforms, and our policies are clear that this includes CSAM."

Third, there are also tools which can literally "nudify" photos of any clothed children on the internet.

And worse, the IWF also said that it is seeing evidence of these AI-generated CSAM being shared and sold online.

In its findings following a month-long investigation, the watchdog said that it found instances of these images on the dark web, a section of the internet that can only be accessed with a specialized browser.

AI-generated CSAM is illegal under the Protection of Children Act 1978, which criminalizes the taking, distribution and possession of an "indecent photograph or pseudo photograph" of a child.

The Coroners and Justice Act 2009, which is an Act of the Parliament of the UK, also criminalizes non-photographic prohibited images of a child, such as cartoons or drawings.

In its investigation, it found at least more than 20,000 AI-generated images posted in one dark web CSAM forum, and selected more than 11,000 for assessment.

"These were the images that were judged most likely to be criminal," the watchdog said in its report.

"The remaining 9,146 AI-generated images either did not contain children or contained children but were clearly non-criminal in nature."

A team consisting 12 IWF analysts were tasked to analyze these images, spending a painstaking 87.5 hours, and concluded that 2,562 of the images were assessed as criminal pseudophotographs, and 416 assessed as criminal prohibited images.

They found that more than half of that depict primary school-aged children (seven to 10 years old), and more than one hundred of them depict children aged three to six, while two images depicted babies (under two years old).

The IWF said the vast majority of the illegal material it had found have breached the Protection of Children Act, with more than one in five of those images classified as category A, which is considered the most serious kind of content, because they can depict rape and sexual torture.

According to Susie Hargreaves, the chief executive of the IWF, in a website post:

Before this, the IWF warned that it had evidence of AI-made abuse emerging.

This time, the IWF said that the trend is getting more apparent, and that it is seeing an acceleration in use of this particular technology.

Hargreaves said that the findings is the watchdog’s "worst nightmares have come true."

The IWF fears that a rise of AI-generated CSAM can distract law enforcement agencies from detecting real abuse and helping victims.

"If we don’t get a grip on this threat, this material threatens to overwhelm the internet," said Hargreaves.