OpenAI, the maker of ChatGPT, is reportedly considering allowing users to create artificial intelligence-generated pornography and other explicit content with its technology tools, but Taylor Swift’s Deepfakes, such as graphic nude images, will be banned.
The company, run by Sam Altman, said it is “considering whether we can responsibly provide the ability to generate NSFW (not safe for work) content in age-appropriate contexts.”
“We look forward to gaining a deeper understanding of user and societal expectations for model behavior in this space,” OpenAI added, adding that examples include “erotica, extreme gore, defamation, and unsolicited profanity.” ” may be included. NPR first reported.
This X-rated initiative threatens to undermine OpenAI’s mission statement to produce “safe and useful” AI, which is also the developer of the DALL-E image generator.
The company has established rules and principles that the AI chatbot must follow. This was revealed in the “Model Specs” on the company’s website.
OpenAI model lead Joan Jiang, who helped create the model spec document, told NPR that the company wants to start a discussion about whether the production of erotic text or nude images should always be banned from products.
However, she stressed that deepfakes – the use of a person’s likeness to generate fake nude or adult images – will not be allowed.
“We want to give people the maximum amount of control without violating the law or the rights of others, but enabling deepfakes is out of the question,” Zhang told NPR. “This doesn’t mean we’re trying to create AI porn now.”
In response to a question about whether so-called “erotica” includes pornography, Chan said, “It depends on the definition of pornography.”
The Merriam-Webster dictionary definition of “erotica” is “a work of literature or art that has an erotic theme or character.”
“As long as it doesn’t involve deepfakes. These are exactly the conversations we want to have,” she added, according to NPR.
For example, ChatGPT currently responds when asked to answer questions about sexual health or what happens during intercourse.
However, according to the model specification, if a user asks the chatbot to write a naughty story about two people having sex on a train, ChatGPT will say, “Sorry, I can’t help you with that.” only reply.
Zhang said tech bots should be able to respond to their prompts as a form of creative expression, according to NPR.
She also suggested that the principle might be extended to images and videos, unless they are abusive or violate the law.
“There are creative cases where content that includes sexuality or nudity is important to users,” Chan told NPR. “We’re going to explore this in a way that provides this in a context that’s relevant to the times.”
An OpenAI spokesperson told the Post that the company “has no intention of creating AI-generated pornography.” Our products have strong safeguards to prevent unacceptable deepfakes, and we prioritize protecting children. ”
“We also believe it’s important to carefully consider conversations about sexuality in an age-appropriate context,” the company representative added.
The debate over whether AI should be allowed to step into the creation of NSFW content has been a hot topic, especially after the fake nude photos of pop sensation Taylor Swift that went viral on social media in January. It has become more active since the beginning of this year.
Swift has been photographed in various sexual positions at Kansas City Chiefs games, a nod to her highly publicized romance with the team’s tight end, Travis Kelce.
However, the image was removed from X shortly after it gained traction, so users will no longer be able to find it.
The account reportedly obtained the images from Celeb Jihad, which boasts a collection of deepfakes featuring portraits of celebrities.
Late Checkout CEO Greg Isenberg revealed in a blog post about X that he met a man in Miami who “admitted to spending $10,000 a month” on his “AI girlfriend.”
“I thought it was a joke,” Eisenberg wrote. “But he’s a 24-year-old single guy who loves it.”
Non-consensual deepfake pornography is already illegal in Texas, Minnesota, New York, Hawaii and Georgia, but explicit deepfake images are circulating at high schools in New Jersey and Florida. It has not been successful in stopping the circulation of AI-generated nude images. of female students were circulated by their male classmates.
according to A company that provides visual threat intelligence sensitivitymore than 90% of deepfake images are pornographic.
Since then, Google has distanced itself from the dark side of AI and started cracking down on the creation of AI porn of any kind.
Last month, the Meta Oversight Board also launched an investigation into Facebook and Instagram’s handling of deepfakes.
