Reality is sometimes stranger than fiction. That’s what a photographer and writer who goes by the name Miles Astray wanted to make clear with his submission to the 1839 Awards, a new photography competition with a category for images created by artificial intelligence. For that AI category, Astray submitted an image that appeared to show a flamingo without its head and neck — but that was a little white lie.
The piece, entitled F L A M I N G O N E, is a real photograph. It was awarded first place by a panel of judges in the AI category and won the People’s Vote Award before it was disqualified.
Astray wasn’t the first person to pull a stunt like this, but such AI trickery usually happens in reverse. Last year, for example, an AI-generated image by artist Boris Eldagsen won the creative image category at the Sony World Photography Awards.
If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
“He kind of did the opposite of what I did,” Astray says of Eldagsen’s submission, “but to send a very similar message: Basically, we’re not really ready for this technology. We’re not really keeping up with how fast it is moving.”
Scientific American spoke with Astray to learn the real backstory of F L A M I N G O N E and explore the question of whether AI can be used to make good art.
About two years ago I was on my way from Canada to Ecuador, and I had a stop in Aruba, kind of like a prolonged layover. And I’d heard about a beach on a little island where flamingos roam around freely. So I got up at five in the morning and took the first boat there.
I’ve seen flamingos in the wild before but always from afar — usually you’re not able to really interact with them. But there [on that island] they don’t seem to be bothered by the presence of humans; they’re not shy at all. And that [flamingo in the photograph] was the first flamingo I came across. It was right in front of me, just going about its morning routine. Then it scratched at its belly, so the head and neck were tucked underneath the torso. I wasn’t aware of it, you know; I was just taking pictures and shooting away. But then, later, I realized that I that I’d gotten a pretty fascinating angle. Of course, a flamingo is already a pretty fascinating and strange creature. But in that moment, it turned into this absolutely surreal creature.
What kinds of images do you normally try to capture?
It’s funny, a lot of photographers have a very particular style that is recognizable, and I don’t really think of myself as that kind of photographer. What is representative of my work, I think, is that it’s candid, that it’s authentic. I’m more of an observer. I really like to depict nature and the world as is. For me, the most fascinating thing is the world itself. My ambition is none other than just to capture it as accurately as I can.
So why did you decide to submit the photograph to an AI image competition?
There have been a few examples of AI imagery being entered into real photo competitions in recent years, and I think the idea of turning that around [and submitting a real photograph to an AI competition] had been like simmering in my head for some time. So then when I saw the flamingo picture, I was like, “Okay, this could really be mistaken for an AI-generated picture.” But there are not that many AI image competitions, so I was kind of waiting for the right moment to come along, and then it presented itself with [the 1839 Awards].
What about the photograph made you think that it could be mistaken for AI?
It’s just so surreal, to the point of looking unreal. A flamingo is already an incredible creature. It looks strange and awkward but also graceful and colorful. And that very moment that I caught there, it’s just this — legged dot, basically.
What did you hope would happen when you submitted the image?
I was hoping that if it won, it would show that nature still outdoes the machine in terms of creativity and beauty and fascination. That was one part of the message. And then the other part was more related to the fact that, right now, we can hardly distinguish between real photos — and even audio and videos — and AI-generated content. And that has its dangers and pitfalls.
[The judges] are not to blame for not catching that this was a real photo. Why would they have expected somebody to enter a real photo into the AI category? It has nothing to do with their judgment or aesthetic sense. I did feel bad about [the deception] to some degree. There are ethical considerations of deceiving a jury and the audience, of course. But I was hoping that they would understand and that they would see the message here. And I was really amazed by [the organizers’] reaction. Of course, they disqualified my entry, which was the right decision — I fully support that. But [the organizers] were so kind. [Editor’s Note: In an e-mail to Scientific American, Lily Fierman, director of the organization that put on the 1839 Awards, said, “We fully appreciate the powerful message Miles relayed with his submission…. We agree that it is an important, relevant, and timely statement.”]
I didn’t expect that response, but they were such good sports about it.
Here, nature outdid the machines. Do you think the machines can ever outdo nature,?
By now, that technology has become so advanced that it can produce those very real-looking images that can compete with [real ones] to some degree. And I think that’s fascinating on one side but also a very slippery slope on the other — because next thing you know, it’s a weapon of mass misinformation. But in the hands of the right person, like an artist like Boris Eldagsen — the images he creates are fascinating. They are incredible scenes. I wouldn’t say they’re more fascinating than reality; I think there’s nothing stranger than reality. But this comes pretty close.
Right now, the technology [for AI-generated images] is still a bit hit-or-miss. For example, last year I did use one of those tools to try it out. I wanted it to replicate one of my favorite pictures that I’ve taken. It’s of a stray cow in the Himalayas in this Rhododendron forest — it’s really a beautiful scene. The results [from the AI] were so cartoonish that I just dropped it right there. But the technology is only going to get better from here, I would imagine. And that becomes a problem once we flood the Internet with that kind of content, and you cannot discern it anymore from real content. I think right now is a very critical moment in time to address all that and get ahead of it.
What are some ways we could address this potential for AI-generated misinformation?
I think one thing we could do is tag AI-generated images and videos [accordingly] so that people have an idea what’s real. And the onus of [mandating] that would be on governments and on the private sector as well. It would be easy enough to put a watermark [on images developed with these AI tools]. But it’s also an individual responsibility to think critically about the things we see. The [younger] generations just starting to navigate the Internet basically won’t know a world without AI, so I think it’s also important that we educate young people how to be critical.
It’s a nuanced matter. I’m not demonizing the technology — it has great potential. I firmly believe that technology itself is not inherently good or bad. It only becomes one thing or another when human beings apply it.