Deepfake victim urges better regulation of 'dangerous'…

Model Georgia Horton is advocating for stronger artificial intelligence regulations after discovering that an acquaintance had used her social media photographs to create non-consensual, sexually explicit deepfake imagery. The incident, which led to a criminal conviction, highlights the severe personal impact and the rapid accessibility of tools capable of generating abusive content.

What happened

Georgia Horton discovered that Ethan Orchard, a former college friend from Hampshire, had manipulated her public social media images to produce pornographic content without her consent. Beyond the creation of these images, Orchard utilized online platforms to share violent and sexually explicit fantasies regarding Horton and other women he knew. He facilitated discussions where participants were asked to choose between victims or describe sexualized violence against them.

During the subsequent police investigation into Orchard’s activities, authorities uncovered a website containing more than one million images of child sexual abuse, illustrating the broader criminal context often associated with these digital platforms. Orchard was ultimately sentenced to three years and four months in prison for his actions, which involved creating sexualized fake images of three different women.

Horton notes that while recent legislation makes the creation of non-consensual sexual deepfakes a criminal offense, current regulations remain insufficient. She argues that the technology is becoming increasingly easy to access and that existing safety features are easily bypassed. She emphasizes that self-regulation by technology companies is failing to prevent the misuse of AI models. Advocating for stricter safety measures and tougher restrictions on software, Horton warns that the widespread availability of these tools empowers predators and poses a dangerous threat to individuals. She specifically urges parents to be cautious about sharing photographs of their children online, noting that a single image can be exploited to generate a limitless supply of harmful, AI-synthesized content.