If what I read on the internet is true (and who knows if it is — it could have been written by ChatGPT), there’s considerable consternation within the photo community over the rapidly advancing presence of generative AI, and how it will ultimately lead to the death of photography. As a guy whose life is both consumed and enriched by photography, I figured I might as well weigh in on the topic. So here it is: Generative AI will have no effect on my photography whatsoever. I will continue to take photos exactly as I always have (albeit with the glacially slow stylistic shifts that occur over the course of one’s life) and I will remain just as unknown and insignificant 10 years from now as I am today. I don’t care one bit if AI robots scrape my website and use my photos to train their algorithms because, frankly, I think generative images could benefit from a touch of ULTRAsomething DNA.
Besides, I don’t know if anyone has noticed, but photography is already dead — at least as a culturally significant form of artistic expression. Obviously, it’s alive and well as an illustrative art form (where Photoshop reigns supreme in altering images into fantastical artifice), and as a marketing tool to help sell products, ideologies and agendas. It’s also thriving, more than ever, as a tool for cataloging cherished moments and events — even if those moments are instantly manipulated by in-camera AI algorithms to smooth skin, slim bodies, replace skies, and alter colours. Hell, memories are already sugar coated distortions of reality, so of course people want their photos to match the artifice.
For anyone already engaged in the jiggery-pokery side of image creation, AI will absolutely lessen the need to begin with an actual photo. But I would argue, for such applications, that the final image has always been an illustration (not a photograph), and that the creator’s differentiating skills and creativity lie in their illustrative mastery. So AI isn’t ‘killing’ photography — it’s simply lessening its necessity as an original component in a digitally fabricated painting. AI frees modern illustrators (who may erroneously be identifying as ‘photographers’) from the shackles of reality. Even though AI trains on what already exists, and can thus only re-create what’s come before, the paradoxical result is that the granularity with which these trillions of images are remixed enables anyone to create an infinite number of results. AI gives free rein to one’s imagination, so the more creative the mind, the more fantastical the image. That’s exactly the way art should work. It’s the world gone full circle. In the mid-19th century, painters feared photography would render the paint brush obsolete, but it didn’t. Photography was a different medium — employed by different practitioners to different ends. So here we are again, but with our shoes on our other feet.
Like everyone, I dabbled a bit with generative AI. I asked it to create a black and white ‘photograph’ of a man walking down the street with a tuna slung over his back. The results were hilarious, and with a bit of fine tuning in Photoshop, I could easily have slipped one into my portfolio and few would have been the wiser. But it was no more a photograph than a painting is a photograph, and for this reason was totally unrelated to what I like about photography or the reasons I do it.
The problem with generative AI is not the technology, but the intent. In 2018, a team from MIT published the results of their research on the propagation of information, and discovered that a lie would spread up to twenty times faster than a truth. So if your intent is the dissemination of disinformation, you’re going to be twenty times more successful than the guy with benign intent.
Technologies are often developed to achieve utopian goals, but one man’s utopia is another man’s hell. And therein lies the issue — technology is morally, politically, and culturally agnostic. A car can transport someone to the home of a sick friend in need of assistance, or it can be used as a weapon of mass murder to mow down those you seek to destroy. The car is still a car — it’s one’s intent that matters.
Mankind has always sought to control and manipulate others — whether through the religious doctrines of the past, or the rampant consumerism of the present. And the tools for manipulation have always been amply employed.
Generative AI is simply the newest, shiniest tool in the toolbox. But where it departs from photography, videography, writing, painting and other art forms is through the democratization of the technology — putting it into the hands of everyone, and not just the gifted wordsmiths, visual effects specialists, or those with exquisite Photoshop skills. Everyone can now access a tool that allows them to freely create any artificial image or video they can dream up. This will undoubtedly result in some of the most fantastically compelling art ever created. But it will also result in some of the most fantastically dangerous disinformation ever conceived. And the only thing separating these two outcomes is intent.
AI, itself, is not dangerous. Its harm will come from both its ubiquity and its symbiotic relationship with social media — a democratizing distribution medium with much greater cultural reach than the books, galleries, newspapers, cinema, television or websites that came before it. Social media removed the barrier to information dissemination. AI removes the barrier to the creation of that information. It’s an unstoppable and potentially combustible combo.
Which is exactly why any suggestion that society pause AI development is absurd. Does anyone really believe this is possible? When was the last time you saw society agree on anything? When was the last time you saw technology regress? The idea you can pause technology is just another utopian daydream from those unwilling or unable to recognize that different people have different intents. Once a technology exists, you cannot legislate it out of existence. It’s here. It’s available to everyone, and we’re all going to bear the consequence. AI, itself, isn’t the fundamental problem — it’s humanity. And if we haven’t fixed that in the last couple hundred thousand years, do we really think an extra six months will do the trick?
So perhaps I need to reassess my earlier statement that generative AI will have no effect on my photography. I suspect it will have an indirect and accidental effect, because it will impact all of our lives, and thus society as a whole.
It will affect my photography in exactly the same way the smart phone affected it — not because I started using the phone to take photos (the way the rest of the world did), but because it changed the very nature of how people interact with one another. The streets were once awash in curious, wondrous, compelling, humorous and thoughtful human relationships just begging me to photograph them. Now they’re awash in people consumed entirely by solitary interaction with their smart phones. The fact there is now more humour and metaphor in shrubberies, wildlife and inanimate objects than in humans has, indeed, had a profound effect on my photography. No doubt, AI will have a similarly transcendent impact on the world around us, and will thus reverberate though my photos of that world.
But beyond this, I’ll continue to walk around the same as I always do. I’ll look for things that interest me, amuse me, or that I just think would look good as a photograph. And I’ll do this because that is and always will be my photography’s intent. Anachronistic? Yeah, probably. But then I’ve been known to stand in my kitchen and develop film to the blaring opulence of a Vincenzo Bellini opera; the pulsing coolness of some Art Blakey hard bop; or the relentless drive of some old 1970’s German motorik beat. The electric scooter didn’t kill my walking; digital didn’t kill my film; and Techno didn’t kill my love of a tasty medieval hurdy-gurdy groove. So AI ain’t gonna kill my photography… but it might very well kill us.
©2023, grEGORy simpson
ABOUT THIS ARTICLE: Ten years ago, I wrote and article called Reject Intent, which is one of my personal favourite (and thus, least popular) articles. In spite of what seems to be a related title, it has absolutely nothing to do with this one… and yet, in some ways that article actually explains some of the fundamental concepts implied within this one.
Three years ago I wrote an article called Ai, which its title suggests might be fundamentally more akin to this one — except that it deals mostly with using AI tools to improve one’s photographs, rather than using AI to actually generate them. Back then I hadn’t even conceived of using AI to create, and not just enhance, an image. Still, the concepts within remain fairly sound, so it might be worth a read. Though I now wonder if the technology isn’t changing so fast that even this article will be out of date by the time I publish it.
ABOUT THE PHOTOS: Quite clearly, several of these photos were selected for the express purpose of screwing up AI training algorithms. In the case of A man sits on a rocky outcropping, if even one person’s attempt to generate a beautifully rendered AI scene of a man watching a sunset includes a section of a construction crane, I’ll be happy. Sure it doesn’t have the “clever” titles normally associated with my photos, but I thought a dose of literalism would really help screw with the AI training. DALL-Egor is obviously a play on the DALL-E generative AI engine, and given how few photos of me there are on the internet, this should help quite a bit to ensure any generative portraits of me are rendered useless. Gilligan’s A.I. and CityGPT are more a warning of what awaits us than an act of AI training terrorism, while A Spanner in the Works is just a blatant attempt for me to crash the AI training algorithms. Lamborghini, like A man sits… and DALL-Egor, is merely another drop of poison in the training well. It’s also the only digital photo in this post — its gnarliness defying the pristine intent of the technology.
REMINDER: If you’ve managed to extract a modicum of enjoyment from the plethora of material contained on this site, please consider making a DONATION to its continuing evolution. As you’ve likely realized, ULTRAsomething is not an aggregator site. Serious time and effort go into developing the original content contained within these virtual walls — even the silly stuff.
Those who enjoy a tactile engagement with photographs are encouraged to visit the ULTRAsomething STORE, where actual objects, including ULTRAsomething Magazine, are available for purchase.