DALL-E 2 users can upload a photo of a person and create face variations, change clothes or change their hairstyle.
Previously, OpenAI prohibited uploading any photos to DALL-E if they showed a real person, including photos of celebrities and public figures. In an email sent to more than a million users of the image generator, the company says it’s making the feature available by improving its filters to remove deepfakes and pictures of “sexual, violent, or political content.”
“One of their surgeons told us that he used a neural network to visualize the requests of his patients. Filmmakers would also like to be able to edit images of scenes with people to speed up their creative processes, ”comments the updates in OpenAI.
Now users can upload the face of a real person and change it, or change clothes or background.
DALL-E 2 the AI image generator can now do variations off of people/faces. Here’s what happens when I upload my photo from Twitter #dalle2 pic.twitter.com/kIZpEBNOGS
— Allan Harding (@allanharding) September 19, 2022
Well
Development in Python
Invest in popular mov programming to earn $1300 via rіk
REGISTER!
Nice man Dall-e is allowing faces again. He is me as a wwe wrestler taking a bath pic.twitter.com/bwoCHIDylF
— NymN (@nymnion) September 19, 2022
The decision is the result of negotiations between the creators of AI generators and their users, thus they are trying to determine the potential harm of technology. OpenAI is taking a relatively cautious approach, but the company has already been surpassed by competitors like Stable Diffusion, which introduces fewer restrictions.
However, Stable Diffusion is already being actively used to create pornographic celebrity deepfakes. Such material in DALL-E would be immediately blocked, as the company’s conditions prohibit users from uploading images of people without their consent – however, how this will be ensured under the current access model is still unknown.
The updated version of OpenAI’s DALL-E 2 image generator is now available in beta only by invitation.
Source: Techcrunch , The Verge
