Meta's open-source AI is being used to create sexbots

The Washington Post spoke to the creator of one such sexbot called Allie. The chatbot was created by a developer who chose to remain anonymous for sexual play.

Similarly, Interesting Engineering reported earlier in May that 23-year-old Snapchat influencer, Caryn Marjorie, was cloned into an AI. Her chatbot often indulges in sex chats and charges $1 per minute.

While many generative AI tools are being used for explicit purposes, Allie is a bit problematic. As per reports, the sexbot also allows users to indulge in graphic rape and abuse fantasies online.

AI models being used for NSFW purposes

Calling it a ‘safe outlet to explore,’ the creator told Washington Post, “Can’t really think of anything safer than a text-based role-play against a computer, with no humans actually involved. It’s rare to have the opportunity to experiment with ‘state of the art’ in any field.” 

He said open-source models allow “people to build products that cater to their preferences without corporate guardrails.” He referred to chatbots like ChatGPT, which are ‘heavily censored’ and can be used in controlled and safe environments.

Amid the growing trend of users turning to AI models like LLaMA to play out their sexual fantasies, industry experts have raised concerns.

Gary Marcus, a cognitive scientist who testified to Congress on AI regulation in May, said, “We don’t open-source nuclear weapons. Current AI is still pretty limited, but things might change.”

Open-source AI models, a boon for developers, are often published on Hugging Face, an AI platform where users upload and share machine learning models.

The company’s CEO, Clem Delangue, in an interview with the Washington Post, acknowledged that tools like Allie could be abused. Still, he also said that the open-source approach allows for more incredible innovation and transparency than closed models.

Hugging Face’s rules allow sexually explicit projects. Still, they have placed restrictions on content that involves minors, or that is “used or created for harassment, bullying, or without explicit consent of the people represented,” reported the Washington Post.