Deepfake porn is sexually explicit content generated by deep learning models that synthesizes a person’s likeness onto an erotic body or scenario. It uses generative adversarial networks (GANs) or diffusion models to produce photorealistic or video imagery that appears to feature the target individual.
Why It Matters
Deepfake porn represents the most invasive application of generative AI to sexuality. Unlike traditional image manipulation, deepfakes can be animated, voiced, and made interactive. The victim’s face is not merely pasted but synthesized — breathing, moving, speaking — creating a visceral illusion of participation in an act that never occurred.
The harm is existential as well as reputational. The victim faces not just embarrassment but a fundamental loss of bodily autonomy. Their image, their most recognizable self, is made to perform against their will. The technology removes the possibility of consent from the visual register entirely.
Example
The first wave of deepfake porn targeted celebrities, with thousands of synthetic videos appearing on dedicated platforms. The second wave has expanded to include private individuals — ex-partners, public figures, anyone whose image is available online. The production requires only a handful of photographs and freely available software.
The AIrotic Angle
AIrotic treats deepfake porn not as a pornography issue but as a consent technology issue. The question is not whether the content is “convincing” but whether the architecture of generative AI makes bodily autonomy technically impossible to protect. The solution is not better detection but better design: systems that refuse to generate likenesses without verification, and legal frameworks that treat likeness as a property right.
Leave a Reply