OpenAI considers allowing production of adult content on ChatGPT

A OpenAIowner of ChatGPT, is considering allowing its artificial intelligence to generate pornographic, violent and explicit content. The assessment was carried out in a document published on Wednesday the 8th, which discusses how OpenAI models should behave.

“We believe that developers and users should have the flexibility to use our services as they see fit, as long as they comply with our usage policies,” said OpenAI, in response to the directive not to allow its models to respond with NSFW (not safe for users) content. work), including erotic content, extreme violence, insults and unsolicited profanity.

“We are exploring whether we can responsibly provide the ability to generate NSFW content in age-appropriate contexts through the API and ChatGPT. We look forward to better understanding user and societal expectations regarding model behavior in this area.” “

Responding to a request from US public radio NPR, Joanne Jang, an employee of the San Francisco-based company who worked on the document, said OpenAI wants to start a discussion about whether the generation of erotic text and nude images should always be banned. of your products. Deepfakes, however, would remain prohibited, according to Jang.

“We want to ensure that people have as much control as possible insofar as it doesn’t violate the law or other people’s rights, but allowing deepfakes is out of the question, period,” Jang said. “That doesn’t mean we’re now trying to create AI porn.”

Jang explains that there are creative cases where content involving sexuality and nudity is important to users.

Through a spokesperson, OpenAI reinforced on Thursday, 9, that the company has no intention of allowing the generation of AI-generated pornography and deepfakes, but that it believes “in the importance of carefully exploring conversations about sexuality in contexts age-appropriate.”

In March, Mira Murati, director of technology at OpenAI, stated that the company does not rule out the generation of nude images on Sorathe startup’s new AI capable of producing realistic videos from text commands.

Pornographic deepfake worries authorities

The use of AI to create nude images of real women, including children and teenagers, has worried institutions and governments around the world. Organizations combating online crime, such as the Internet Watch Foundation (IWF), have been warning about the use of AI tools to manipulate photos of children and teenagers published on the internet to make them nude. With the photos in hand, criminals can blackmail the person into sending more photos with compromising content. The situation, however, has also increased among adolescent and adult women.

Explicit images created by deepfake of singer Taylor Swift were widely publicized in X (formerly Twitter) at the beginning of the year, with one of the photos gaining tens of millions of views and 24 thousand reposts. After a collective effort by the artist’s millions of fans, which lasted several days, the network started to ban the material.

In Brazil, the police investigate numerous cases, in different Brazilian states, of students who report the disclosure of nudes fake created with artificial intelligence.

In April, Apple and Google removed apps that promised nude deepfakes from the App Store and Google Play, following complaints from 404 Mediawhich reported that Meta was hosting ads for the apps on Facebook and Instagram.

According to Meta, around 1,500 advertisements for “AI Girlfriends” appeared on the company’s platforms in the US by April 30, 2024.

In Brazil, Bill 5694/23 establishes a sentence of four to eight years in prison, as well as a fine, for anyone who uses artificial intelligence to manipulate images and sounds of children and adolescents with the aim of producing sexual content. The text changes the Child and Adolescent Statute and is under analysis in the Chamber of Deputies.