We’ve already seen stories about AI images being sent to photography and arts competitions. The algorithms have become so prolific that in some instances, the result is nigh indistinguishable from real man-made art.
But photographer Miles Astray had a different idea.
He submitted a real photography to an AI competition — the 1839 Color Photography Awards.
“I wanted to show that nature can still beat the machine and that there is still merit in real work from real creatives,” Astray tells PetaPixel over email. “After seeing recent instances of AI-generated imagery beating actual photos in competitions, I started thinking about turning the story and its implications around by submitting a real photo into an AI competition.”
The picture is a pretty surreal (but very real) image of a flamingo, hiding its head away. At first glance, you could be tricked into thinking it’s a messed-up AI image. And that’s probably also what tricked the competition judges as well — because the photo won 3rd prize in the contest. It even won the People’s Vote award.
We’re still really bad at recognizing AI images
If you’ve ever felt bad about being tricked by an AI-made image, well, it happens to everyone. The judges for this AI photography contest worked at the New York Times and Getty Images, among others. Apparently, no one figured out that the photo was real.
Astray said his experiment was successful.
“Of course, I feel bad about leading the jury astray, but I think that they are professionals who might find that this jab at AI and its ethical implications outweighs the ethical implications of deceiving the viewer, which, of course is ironic because that is what AI does.”
“I’m glad to see that this experiment confirmed my hypothesis: there is nothing more fantastic and creative than Mother Nature herself,” said Miles Astray who has no ambition to work with AI. “I don’t demonize the new technology and see its potential, but currently I see its limitations and dangers even more clearly.”
The competition organizers said they also empathize with the message of the entry, but that they still have to disqualify it.
“No one believes in the power of photography more than we do. We’ve asked Miles to work with us and give a statement for a future blog post we are going to do regarding this exact topic, with his submission, press release, and any statement he shares as a jumping-off point. As an artist, his voice will make a difference in this conversation. However, after much internal debate, we decided to disqualify his entry into the AI category in consideration of the other artists who submitted their work. Our contest categories are specifically defined to ensure fairness and clarity for all participants. Each category has distinct criteria that entrants’ images must meet.”
So where does this leave us?
We still have no reliable way of telling apart AI photography and real images. Our eyes and brains are not trustworthy for this task and other algorithms are also pretty unreliable. At the same time, we’re starting to run into AI-generated imagery more and more.
Art and photography competitions are one thing — they’re supposed to make us think. But what happens when we have AI images depicting politicians, or even our colleagues? For now, only a few algorithms can create images realistic enough to trick people, and they have safeguards, but other companies aren’t so ethical, and soon enough, we’ll be facing a deluge a of AI-generated images in our daily lives. The implications are vast and multifaceted, affecting trust, authenticity, and our perception of reality.
Nature may be unparalleled, but technology is already good enough to fool us. Who knows what implications this will have?
