Julia, 21, has received fake nude photos of herself generated by artificial intelligence. The phenomenon is exploding.
“I’d already heard about deepfakes and deepnudes (…) but I wasn’t really aware of it until it happened to me. It was a slightly anecdotal event that happened in other people’s lives, but it wouldn’t happen in mine”, thought Julia, a 21-year-old Belgian marketing student and semi-professional model.
At the end of September 2023, she received an email from an anonymous author. Subject: "Realistic? “We wonder which photo would best resemble you”, she reads.
Attached were five photos of her.
In the original content, posted on her social networks, Julia poses dressed. In front of her eyes are the same photos. Only this time, Julia is completely naked.
Julia has never posed naked. She never took these photos. The Belgian model realises that she has been the victim of a deepfake.
Why not blame the spread? You can’t ban the tool, it’s easily accessible software and that only requires easily accessible consumer hardware, and you can even semi easily train your own models using easily accessible porn on the Internet, so if you want to ban it outright, you’d need to ban the general purpose tool, all porn, and the knowledge to train image generation models. If you mean ban the online apps that sell the service on the cloud, I can get behind that, it would increase the bar to create them a little, but that is far from a solution.
But, we already have laws against revenge porn and Internet harassment. I think the better and more feasible approach that doesn’t have far reaching free speech implications would be to simply put heavy penalties on spreading nudes images of people against their will, whether those images are real or fake. It’s harassment as revenge porn, and I didn’t see how it’s different if it’s a realistic fake. If there is major punishment for spreading these images then I think that will take care of discouraging the spread of the images for the vast majority of people.