On the Role of Artificial Intelligence in Disseminating Harassment, Blackmailing, and Embarking in Deep Fake Pornography
AI-generated pornography has quickly become one of the biggest and most troubling applications of the type of generative AI technology OpenAI has pioneered. Deepfake porn, which is porn made with artificial intelligence and features fake people without their consent, is a common tool of harassment against women and girls. In March, WIRED reported on what appear to be the first US minors arrested for distributing AI-generated nudes without consent, after Florida police charged two teenage boys for making images depicting fellow middle school students.
Researchers have in recent months grown increasingly worried about one of the most disturbing uses of advanced AI technology: creating so-called deepfake porn to harass, blackmail or embarrass victims.
The note says that they are looking into whether or not they can ethically provide the ability to generate content in age-appropriate contexts. User and societal expectations are expected to be better understood in this area.
“If we want to ensure that people have control over it, we need to make sure it does not violate the law or other peoples’ rights”, Jang said. “This doesn’t mean that we are trying now to create AI porn.”
While Jang stresses that starting a debate about OpenAI re-evaluating its NSFW policy does not necessarily suggest drastic rule changes are afoot, the discussion comes during a fraught moment for the proliferation of harmful AI images.
A University of Virginia School of Law professor says that intimate privacy violations, including sex videos and other nonconsensual synthesized intimate images, are a serious problem. “We now have clear empirical support showing that such abuse costs targeted individuals crucial opportunities, including to work, speak, and be physically safe.”
Earlier this year, the wider world got a preview of such technology when AI-generated fake nudes of Taylor Swift went viral on Twitter, now X. Microsoft added new safeguards to its text-to-images artificial intelligence generator, according to the tech news publication.
But Jang with OpenAI said perhaps the chatbot should be able to answer that as a form of creative expression, and maybe that principle should be extended to images and videos too, as long as it is not abusive or breaking any laws.
Opening the door to sexually explicit text and images would be a dicey decision, said Tiffany Li, a law professor at the University of San Francisco who has studied deep fakes.
“The harm may outweigh the benefit,” Li said. “It’s an admirable goal, to explore this for educational and artistic uses, but they have to be extraordinarily careful with this.”
Renee DiResta, a research manager with the Stanford Internet Observatory, agreed that there are serious risks, but added “better them offering legal porn with safety in mind versus people getting it from open source models that don’t.”
Allowing for any type of computer-generated image or video porn would be used by bad actors to cause the most damage, according to Li.
Text-based abuse can be harmful but it’s not as bad as the other way around. “Maybe it can be used in a romance scam. That might be a problem.
OpenAI Model Spec: A Project to Bring More Transparency to the OpenAI Platform and to the public via Explicit Content Generation
According to Grace McGuire of OpenAI, the Model Spec was an attempt to bring more transparency to the development process, and to get a cross section of perspectives and feedback from the public. She did not give any information about OpenAI’s exploration of explicit content generation or how the company received feedback on the idea.