Jason Allen used Midjourney to create AI-generated art. It came in first at the Colorado State Fair fine arts competition art competition. The intense reactions were immediate.
In Jason Allen’s tweet announcing his win, he describes how he fine-tuned his word prompts. Then he upscaled the art, printed it on canvas and entered it into the Digital Arts / Manipulated Photography category. He added, “I’ve set out to make a statement using Midjourney in a competitive manner and wow! I could not be more excited about having won with my favorite piece.”
Is AI-generated art actually art?
People wasted no time in their reactions. All the way from extreme negative reactions about “vapid art movements” to “the finished product is what matters.” One of the funnier comments mentioned that “AI artwork is the ‘banana taped to the wall’ of the digital world now.”
As mentioned above, it was entered into the Digital Arts / Manipulated Photography category. Several reports also indicate that Allen submitted as “Jason Allen via Midjourney.” Whether the judges knew that this meant it was AI-generated art, only the judges can say.
What does this mean for photographers?
To be sure, whatever your opinions may be, AI-generated art blurs the line between photography and digital arts. The very title of the art category offers a further clue.
Keying in word prompts to have something create AI-generated art already makes many photographers stand up and pay attention. History-minded photographers can recall that not so long ago, art aficionados often said that simply “capturing reality” does not make photography “art.” For many photographers, this debate might hit close to home.
The death of photography?
But DALL-E 2 from OpenAI takes this to another level. Now, DALL-E 2 can “make realistic edits to existing images from a natural language caption. It can add and remove elements while taking shadows, reflections, and textures into account.” DALL-E 2 can also take an image and create endless variations of it.
A user can select an area, type in word prompts, and change the picture on a wall to something else. If you wish, you can add more petals to a flower. You can remove a blurry part of an image and make it sharp, all through word prompts. You can add a centipede wearing a top hat and army boots dancing on a marble.
Oh sure, you can do much of this with Photoshop. But here, your word prompts are guiding easy, near-instantaneous changes that seem limitless. And the results are surprisingly impressive even at this beta-stage of DALL-E 2’s development. There’s nothing to stop you from manipulating your photos to where they are not recognizable from where you began.
What role do we have in the future?
What role will photographers play in the future if you can manipulate your images to such a degree? What role will photographers or artists play if you can generate realistic images of whatever they want? This feels like it could be the death knell for certain genres of photography such as stock photography.
Many artists are pissed. And can you blame them? If you have studied all my life to create art only to have someone enter key prompts and take first prize, how would that make you feel? But pissed-off artists have never stopped the march of technology. AI-generated art — or AI-generated photo editing — will continue whether we like it or not. Stay pissed or adapt, the choice is ours.
The erosion of trust
Our audience already has a deep mistrust for photography. Already, “Photoshop” is used as a verb: “I wonder if that’s ‘Photoshopped?'”. In the not-so-distant future, we might say, “Not sure if it’s real … that could be totally Midjourneyed!” It’s like this with news, science, and videos as well. Why would photography be spared? I’m not sure anyone believes “the camera never lies” anymore.
Why is AI-generated art such a hot thing now?
In a word, accessibility.
Some months back, the Wombo Dream app made a huge splash. Many of us began creating art. We put together unlikely and often milk-out-your-nose funny word prompts to create fascinating and often hilarious images.
DALL-E 2 and Midjourney have offered the participants the opportunity to sign up as beta testers, the latter through Discord.
Last year, when I first began dabbling with AI-generated art, it was tweaky. It often involved adding complex (to me, anyway) computer instructions in addition to the word prompts. It was confusing and required a steep learning curve and some knowledge of code. It’s much easier now, and the quality of the art has evolved tremendously.
I’ve left mostly open-ended questions at this point. Certainly, one aim of an article like this is to provoke thought and opinions, not just tell my own.
I’m excited about the possibilities of AI-generated art, and yes, AI-generated photo editing. I have experimented with the former myself.
However, my concern has not wavered. In 2020, I wrote, “AI is in its infancy. But we have already seen the startlingly real “deepfake” videos and how, it’s difficult to tell reality from fiction. How will this impact social media? How will this impact a public that already has struggles with discerning “fake news” from scientific fact?”
We’ve also discussed AI-generated art on the Photofocus Roundtable podcast. We touched upon people’s trust with the medium.
And although this is hardly my most pressing concern, as a night photographer, I frequently encounter people assuming that the very surreal long-exposure images with light painting are “Photoshopped.” How sad is it that you spend years learning your craft and spend endless hours driving to locations only to find that people think you “faked it?”