In March, the Sony World Photography Awards announced the creative photography category winners. It is a black and white image of an elderly woman hugging a young woman entitled “PSEUDOMNESIA: The Electrician”. A press release announcing the award described the photograph as “haunting” and “recalling the visual language of 1940s family portraits.”
However, Berlin-based artist Boris Eldersen declined the award. He announced that his photo was not a photo at all. He did so with the creative prompts of his DALL-E 2, an artificial intelligence image generator.
“I applied as a cheeky monkey. [competitions] AI images are ready to enter. It’s not,” Eldagsen explained on his website. His stunt has sparked controversy and conversation about when AI-generated or assisted imagery should be considered art.
Scientific American We spoke with Eldagsen about image creation and the future of AI-powered “promptography.”
[An edited transcript of the interview follows.]
What made you start AI art?
I started doing photography because painting was a lonely job. I was always experimenting. So when AI Generator started, I was hooked from the start. For me as an artist, an AI generator is absolute freedom. It’s the kind of tool I’ve always wanted. As a photographer I have always worked from the imagination, but now the material I work with is knowledge. And if you’re older, that’s a plus. If I was 15, I probably would have produced Batman.
Where did the inspiration for “The Electrician” come from?
I did it myself as an exercise and I love the results. It started with a project that started a few years ago. My father he was born in 1924. When he was 17 he went to war, but like many of his German contemporaries, he never spoke of the war. After his death, I found some images of him in his 40s that his mother and I had never seen before. Learned a lot about their time period just by looking at these images. .
The first experiment was to see if AI could be used to reproduce images from that time. And then the “electrician” grew up. The best images are the ones that never occurred to you. They got out of the process. Once started, it will lead you somewhere. The same is true for AI. Start somewhere and make different decisions. Remove elements and add frames. Sometimes AI makes very good suggestions. Sometimes it’s just crap. It takes time and patience, so it won’t be over in 20 seconds. It can take days.
So how did we actually create this image?
I used DALL-E 2 and it was all done by text prompt and repair and repair. For restoration, you can say, “I don’t like his tie,” cross it out and write, “I want him to wear a white tie.” Then get suggestions. And if you don’t like any of these suggestions, start over.out painting [is what] Do if the frame is not big enough. Add a frame so you can see the entire tie, pants, chair, and floor. it is infinite.
Why did you decide to enter the photo contest?
I have been deeply involved in AI and photography. I became one of his German experts. I wanted to test whether the contest allows for the submission of AI-generated images. I submitted to his 3 different contests and the image was always a finalist. there is something in the image. When I apply, I don’t say it was generated by AI. The information is very short. Just an image and a title. When it was picked, I said the art was generated by an AI.
What I wanted happened: a conversation started and it was basically with the help of the community. I thought it would be a week long conversation on the European photo scene.
Can you tell just by looking at an image that someone was AI-generated?
of course. There is a difference in color due to repainting. The left side is too yellow, the right side changes to black and white. And you can probably see not only the fingers, but also part of the right arm.
Have you ever been fooled by AI images?
There is a German magazine called Geo; it is something like national geographic into German. They did an online test with the image and asked, “Is it generated or is it real?” I failed once.
As for “The Electrician”, it’s an old image from early September last year, so I think it’s very easy to understand. But by the end of this year, I don’t think we can say.
Does it surprise you?
As an artist, I love it. As a citizen, I am very worried. Most types of photography can be augmented by AI, but not the photojournalistic part. The media needs to come up with a system that makes it clear what is authentic, manipulated and generated. His AI-generated photos of Pope Balenciaga should have always been annotated. Otherwise, democracy will be manipulated and misinformed by those who can write five words.
But fact-checking is hard work. It takes time. So who is going to pay for all of the staff to do that, to support AI technology? Now, as a citizen, I can’t leave the media alone I say no.it is very important for a democratic society [to be able to distinguish real photos from fakes]Therefore, we have to think about co-financing mechanisms. [fact-checking] As a citizen, as a democratic nation. But how can we co-fund and preserve press freedom? This is something we have to think about.
Aside from the future of democracy and journalism, how will AI fit into the art world?
One thing I would suggest is to clean up the terminology and avoid calling realistic AI art “AI photography”. Because this is not a photo. One of the suggestions that came out of the community was “Promptography” and I love it. Large enough that the result looks like a painting, like a painting, like a photograph.
The next step is to talk about the relationship between promptography and photography. Do they belong to one basket, one museum, one festival, one gallery, one competition? And I don’t have an answer for that. The only thing I can say is that for those of you who want to go back to the analog age, and those who say propography is photography, the short answer is nonsense. You have to think deeper than that.
