In a landmark case highlighting the misuse of AI in criminal activities, a child psychiatrist from North Carolina, David Tatum, has been sentenced to a 40-year prison term.
David Tatum, 41, was found guilty of one count of production of child pornography, one count of transportation of child pornography, and one count of possession of child pornography.
However, what sets this case apart is his use of AI to digitally manipulate clothed images of minors, rendering them sexually explicit. Tatum illicitly recorded minors while they were undressing and showering and used AI to edit them into sexual images.
The US Attorney’s Office for the Western District of North Carolina confirmed the AI-related nature of these offenses, noting the severity of the crime. The case is also particularly disturbing as some of Tatum’s victims were reportedly his own patients.
US Attorney Dena J. King, in her statement, highlighted the gravity of Tatum’s actions: “As a child psychiatrist, Tatum knew the damaging, long-lasting impact sexual exploitation has on the wellbeing of victimized children. Regardless, he engaged in the depraved practice of using secret recordings of his victims to create illicit images and videos of them.”
King further emphasized the misuse of AI, stating, “Tatum also misused artificial intelligence in the worst possible way: to victimize children.”
She also reassured that her office is dedicated to prosecuting those who exploit such technologies for harming children.
The sentence handed down to Tatum includes 40 years in prison and an additional three decades of supervised release.
FBI special agent Robert M. DeWitt expressed his shock at the revelations: “It is horrific to believe anyone would secretly record children undressing and showering for their own sexual gratification.”
There has been high-profile debate surrounding how the law interacts with AI-generated or modified illicit imagery, with US state prosecutors banding together to pressure Congress on producing Federal laws in September.
There have been two incidents in schools, one in New Jersey and one in Spain, of minors being in some way impacted by AI-generated sexually explicit images.
This case, seemingly the first of its kind, should convey that such crimes are not permissible via AI interaction and will be handled proportionately.
Not an isolated incident
A shocking case, this sadly isn’t an isolated incident but part of a worrying global trend in the misuse of AI technology for creating deep fake images of child sexual abuse.
According to a report by the Internet Watch Foundation (IWF), there has been a significant rise in AI-generated images of child abuse, particularly on the dark web.
The UK-based watchdog identified nearly 3,000 such images that violated UK laws. These AI models are being fed with images of actual abuse victims to generate new abusive content. They are also used to create images of underage celebrities in abusive scenarios and ‘nudify’ photos of clothed children found online.
The IWF has expressed grave concerns about this trend, with Chief Executive Susie Hargreaves stating that their “worst nightmares have come true,” in reference to the acceleration of AI technology for malicious purposes.
In the foundation’s month-long investigation of a dark web forum, the IWF revealed that a significant portion of the images breached UK law, with 20% categorized as the most severe form of abusive content.
Easy access to open-source AI models like Stable Diffusion, which don’t have the same safeguards as commercial models, exacerbates the problem, according to the IWF.
