Ai-generated Child Sexual Abuse Images Threaten the Internet, Urgent Action Needed, Warns Watchdog

Ai-generated Child Sexual Abuse Images Threaten the Internet, Urgent Action Needed, Warns Watchdog

Ai-generated child sexual abuse images have the potential to inundate the Internet if proactive measures are not taken, warns a watchdog organization. The existing alarming surge in child sexual abuse imagery on the Internet could escalate further if artificial intelligence tools that generate deepfake photos are not subjected to stringent controls, cautions the UK-based Internet Watch Foundation (IWF) in a recent report. 

The IWF urges governments and technology providers to act swiftly before law enforcement investigators are overwhelmed by a flood of AI-generated child sexual abuse images, leading to a significant increase in the number of potential victims. The chief technology officer of the watchdog group, Dan Sexton, emphasizes that this is not a hypothetical concern but an ongoing issue that demands immediate attention.

The IWF conducted an investigation into more than 11,000 images shared on dark web forums within a span of one month. Shockingly, over 2,500 of these images portrayed child abuse, with one in every five images classified as Category A, representing the most severe forms of abuse. In a notable legal case in South Korea, a man was sentenced to two and a half years in prison for using artificial intelligence to produce 360 virtual child abuse images, setting a precedent for the criminal application of AI technology. 

Disturbingly, there have also been instances where children themselves have been using these tools to exploit each other, as exemplified by an ongoing police investigation in southwestern Spain, where teenagers allegedly used a phone app to manipulate fully dressed pictures of their schoolmates to make them appear nude.

The report sheds light on the dark side of the race to develop generative AI systems that allow users to describe their desired output in words, ranging from emails to artwork or videos, and have the system generate it accordingly. If left unchecked, the deluge of dee

A group of new AI image-generators made a strong impact last year, impressing the public with their ability to produce captivating and lifelike images instantly. However, these cutting-edge tools are not preferred by producers of child sexual abuse material due to their built-in mechanisms to block such content.

According to Sexton, technology providers that offer closed AI models and have full control over their training and usage, like OpenAI's DALL-E image-generator, have been more effective at preventing misuse.

In contrast, an open-source tool called Stable Diffusion, developed by London-based startup Stability AI, has become favored by producers of child sexual abuse imagery. When Stable Diffusion emerged in the summer of 2022, a subset of users quickly learned to exploit its features to generate explicit content, including nudity and pornography. Although most of the material depicted adults, it often involved nonconsensual scenarios, such as the creation of nude images inspired by celebrities.

Stability AI later introduced new filters to block unsafe and inappropriate content, and the software is accompanied by a license that explicitly prohibits illegal use.

In a statement released on Tuesday, Stability AI stated its strong disapproval of any misuse for illegal or immoral purposes across its platforms. The company emphasized its resolute support for law enforcement in combating those who exploit their products for illegal or nefarious ends.

However, it is worth noting that unfiltered older versions of Stable Diffusion are still accessible, and these legacy versions overwhelmingly attract users engaged in creating explicit content involving minors, according to David Thiel, chief technologist of the Stanford Internet Observatory, another organization monitoring the issue.

Sexton remarked," You ca n’t regulate what people are doing on their computers, in their bedrooms. It’s not possible." He questioned how the availability of openly accessible software can be confined to help its use in creating dangerous content like this.

While AI-generated child sexual abuse images would generally be considered illegal under existing laws in the US, UK, and other jurisdictions, the question remains whether law enforcement possesses the necessary tools to effectively combat the problem.

The release of the IWF's report is strategically timed ahead of a global AI safety gathering scheduled for next week, which will be hosted by the British government and feature notable participants such as US Vice President Kamala Harris and leaders from the technology sector.

"I am optimistic even though this report presents a concerning situation," remarked Susie Hargreaves, CEO of the IWF, in a prepared written statement. She stressed the importance of raising awareness about the darker side of this remarkable technology and initiating discussions across a broad audience.