OpenAI Begins Allowing Users To Edit Faces With DALL-E 2 (techcrunch.com) 17
An anonymous reader quotes a report from TechCrunch: After initially disabling the capability, OpenAI today announced that customers with access to DALL-E 2 can upload people's faces to edit them using the AI-powered image-generating system. Previously, OpenAI only allowed users to work with and share photorealistic faces and banned the uploading of any photo that might depict a real person, including photos of prominent celebrities and public figures. OpenAI claims that improvements to its safety system made the face-editing feature possible by "minimizing the potential of harm" from deepfakes as well as attempts to create sexual, political and violent content.
In an email to customers, the company wrote: "Many of you have told us that you miss using DALL-E to dream up outfits and hairstyles on yourselves and edit the backgrounds of family photos. A reconstructive surgeon told us that he'd been using DALL-E to help his patients visualize results. And filmmakers have told us that they want to be able to edit images of scenes with people to help speed up their creative processes [We] built new detection and response techniques to stop misuse."
The change in policy isn't opening the floodgates necessarily. OpenAI's terms of service will continue to prohibit uploading pictures of people without their consent or images that users don't have the rights to -- although it's not clear how consistent the company's historically been about enforcing those policies. In any case, it'll be a true test of OpenAI's filtering technology, which some customers in the past have complained about being overzealous and somewhat inaccurate. Deepfakes come in many flavors, from fake vacation photos to presidents of war-torn countries. Accounting for every emerging form of abuse will be a never-ending battle, in some cases with very high stakes.
In an email to customers, the company wrote: "Many of you have told us that you miss using DALL-E to dream up outfits and hairstyles on yourselves and edit the backgrounds of family photos. A reconstructive surgeon told us that he'd been using DALL-E to help his patients visualize results. And filmmakers have told us that they want to be able to edit images of scenes with people to help speed up their creative processes [We] built new detection and response techniques to stop misuse."
The change in policy isn't opening the floodgates necessarily. OpenAI's terms of service will continue to prohibit uploading pictures of people without their consent or images that users don't have the rights to -- although it's not clear how consistent the company's historically been about enforcing those policies. In any case, it'll be a true test of OpenAI's filtering technology, which some customers in the past have complained about being overzealous and somewhat inaccurate. Deepfakes come in many flavors, from fake vacation photos to presidents of war-torn countries. Accounting for every emerging form of abuse will be a never-ending battle, in some cases with very high stakes.
= They either do it, or somebody else gets ahead (Score:2)
I left DALL-E way long ago. (Score:2)
The irony is that "OpenAI" isn't at all "open", while StableDiffusion is. StableDiffusion is Linux to DALL-E's Windows. Run it on your graphics card. Run it on the cloud and just pay for your compute time. The only license terms on what you create sum up to "Do No Evil", with no software controls to try to "force" you, just left up to your discretion. The software is all open source; change it to your heart's content (and people have been, with seemingly a new branch forming every day and the main branch
Re: (Score:2)
Best thing about Stable Diffusion (Score:5, Interesting)
The best thing about Stable Diffusion is the competition it enforces.
It's a pain to get working, it's even more painful to get good images of people out of it. But, even the possibility of allowing people to assemble a workflow that allows for correcting of faces (and bodies), opens the door to force the rest of the services to enable those features.
I've generated some cool stuff with stable diffusion and midjourney. I'm attempting to assemble my own functional toolset for some particular gaming asset functionality.
So far this has been quite fun.
Re: Best thing about Stable Diffusion (Score:2)
Yup, they can talk about all the people who contacted them about "good" use cases all day long, but I don't for a moment think they would have changed their minds, if it wasn't for the arguably superior model which doesn't have the restriction.
Loab (Score:2)
They've learned about Loab and decided this is the only way to stop her before she manifests completely in our world and destroys everything.
Re:Loab (Score:4, Informative)
For those not in the know, this is Loab. [twitter.com]
Now you know her.
"Allows" (Score:4, Insightful)
OpenAI is a disgrace as nothing about it is open. You can't run your own instance and you require permission just to use the damn thing. If they were being honest then they would change their name to OurAI.
Get a life (Score:1)
Get a life. Your man-love is obvious and no one cares, loser.
Links I absolutely will not click (Score:2)
"A reconstructive surgeon told us that he'd been using DALL-E to help his patients visualize results"
--closes tab, runs out of room
I'm confused (Score:2)
any photo that might depict a real person, including photos of prominent celebrities and public figures
When did prominent celebrities and public figures become real people?
fuck openai and dalle 2 (Score:1)