DALL-E 2 is artificial intelligence software that can turn anything you type into art, in any style. Want a Renoir-style portrait of a panda? Here we go!
A sugar fairy eating a Kehinde Wiley-style cheeseburger? Try this! One woman said: “This is both scary and fascinating!”
People use DALL-E to make music videos…
… as well as children’s books and magazine covers.
I’ve even used it to illustrate “Sunday morning” stories.
DALL-E 2 and its competitors like Midjourney and Stable Diffusion are available for everyone; They are inexpensive or even free to use. It’s easy to see how this technology will change the game in graphic design, interior design, architecture, fashion and filmmaking.
DALL-E creator Aditya Ramesh works at OpenAI, a company founded in 2015 by Elon Musk and others. “The goal of the company is to develop artificial general intelligence,” he said. “And by that we mean an AI that can do all the things that a human can and use them in a way that is safe and maximizes positive benefit to society.”
To train DALL-E (the name is a cross between Dali the artist and Wall-E the Pixar robot), the company fed him 600 million annotated images from the internet.
“It’s not just cutting and piecing; his understanding of images is more conceptual and abstract, much like a person would get inspiration from all the images he or she has seen in their lifetime,” Ramesh said.
I know what you’re thinking: This will put many artists out of work.
Meet concept artist Karla Ortiz, who designed characters, creatures and costumes for many Marvel films including Doctor Strange. “Why would anyone hire anyone when they could just get something? [AI] is that ‘good enough’?” she said.
But their biggest concern isn’t unemployment; It is that the work of professional artists ended up in the OpenAI database. DALL-E knows how to imitate the style of Norman Rockwell or Picasso or Ansel Adams or living, working artists.
“It’s an invasion of privacy,” Ortiz said. “It’s an intrusion into our ability to consent to hacking into those records. No one asked us. The way to fix this is to create records fully populated with public domain works, and then any other kind of extension of it through license agreements.
The idea is if an artist “signs up” to the AI’s image database and becomes part of their algorithms, the artist is compensated.
The AI art companies say they are listening. For example, Stability AI recently announced that it will allow artists to opt out of future versions of their database.
But OpenAI also worried about other downsides, like people churning out AI-generated images with porn, violence, or misinformation. Ramesh said: “When we trained the model, we filtered out images of guns, blood and gore.”
So if someone typed the phrase “The President kills kittens” into DALL-E’s image generator, it would return an error message. “He won’t let that happen,” Ramesh said.
DALL-E also seeks to balance racial and gender stereotypes in Internet imagery. Although 90% of doctor pictures on the internet may be white males, Ramesh said DALL-E would try to balance things out.
But not all AI companies have such safeguards built in. According to Stability AI CEO Emad Mostaque, “It deserves a lot of different views and perspectives on it. And as a society, we need to come together and figure out how best to leverage this amazing technology.”
Stability AI’s Stable Diffusion program is open source, meaning it’s free for everyone, with no restrictions or guard rails. This approach has rung many alarm bells.
Mostaque said, “We think it’s better to make this public, so people can see the power of the technology and then figure out together how we can mitigate the damage, than it being in the province of unelected companies.”
Some Stable Diffusion fans produce harmful and shocking images, but according to Mostaque, they rarely see the light of day because Twitter and Facebook block them out. “If you put it on social media or post it there, it will be treated like any other bad content,” he said.
The state of the art in AI art is getting there. These early apps still have issues with text, faces, and generating the usual number of fingers.
But they are improving quickly. And in the meantime, AI apps that generate audio and video are already being tested.
For the artist Karla Ortiz, these are depressing developments. She believes there is value in the creative process itself: “It’s therapeutic. It’s inspirational. It communicates between people. AI tools cannot do that yet.”
However, Emad Mostaque, CEO of Stability AI, is totally reliant on the AI arts. “I think it’s one of the biggest advances we’ve made in technology since maybe the internet,” he said. “It will create brand new industries and make media even more exciting and entertaining. I think that creates a lot of new jobs.
“It’s inevitable. And I think it’s just going to change everything.”
For more information:
Story produced by Sara Kugel. Publisher: George Pozderec.