Any time new technology emerges in any field of endeavor, there are alway worries and fears about what the technology can do, what it will do, and the effect it is going to have on the status quo. That’s normal. It’s human nature to question, assume the worst, and feel threatened by a new technology that is not understood and its benefits are unknown. We don’t like change, especially sudden change, and imagery created through the use of artificial intelligence, or AI, has unquestionably been a sudden change.
AI imagery has been with us in some form since at least the 1990s, perhaps even earlier. But, it wasn’t until the January 2021 introduction of DALL-E, the first text-to-image generator, that things really took off. Since then, every month has brought further advancement and evolution in what the software can do. There seems to be no end in sight as these images are rapidly becoming more and more realistic with every passing day. Photographers have certainly taken notice and are expressing their concerns with this technology.
“From today, painting is dead.” – Paul Delaroche, French painter
In approximately 1840, French painter Paul Delaroche saw the first daguerreotype image and declared the end of painting. I know painting and photography are not the same. But, I can’t help but think the feelings being expressed now about the future of photography are very similar to the feelings Mr. Delaroche expressed about the future of painting.
There is no doubt that AI images will be disruptive to photography, especially for those for whom photography is their primary source of income. It’s possible and, in some cases, even likely that some photographic genres may cease to exist as we know them today. But, photography has experienced disruptions before, and the secret to surviving those disruptions has been adaptability.
When digital cameras came onto the scene, film photographers, which was every photographer, decried the end of photography. But, photographers who adapted to the new technology survived while those who didn’t struggled. Yes, film photography still exists, but it has been left behind and will never regain its prominence.
The emergence of microstock agencies in the photo licensing market presented a significant challenge to photographers who earned a big part of their income through licensing their images. Again, the photography world adapted with more photographers offering photography tours, workshops, and seminars now than I saw twenty years ago.
What separates AI imagery from photography? The human element. I’ve seen some amazing AI images, but every one I have seen is lacking the emotional content of a photograph. An AI image is an image with no feeling, no connection to the human experience. Why are photography tours and workshop so popular? Because of the human element. Participants enjoy being with like-minded people who enjoy photography. They want to learn new techniques or, perhaps, just learn photography. They want to travel to new locations. They want to share new experiences, make new friends, and have stories to tell when they return home. Those are things AI simply cannot offer.
There is no question that AI imagery will have an impact on the creation of images, and that is certainly worrisome, perhaps even frightening. But, if we continue to focus on the factors that distinguish a true photograph from an AI image, new opportunities are sure to arise, and those who are willing and able to adapt and take advantage of those opportunities will survive.
Photography: the art or process of producing images by the action of radiant energy and especially light on a sensitive surface (such as film or an optical sensor) – Merriam-Webster online dictionary
Art: the conscious use of skill and creative imagination especially in the production of aesthetic objects – Merriam-Webster online Dictionary
Are AI images photography? Are they even art? I’ve seen this question come up online several times, and the usual response is no to both. I agree that AI images are not photography. A photograph is produced by the recording of light on a light-sensitive medium, such as a camera sensor. That isn’t how an AI image is created, therefore, it isn’t a photograph.
But, the declaration that AI imagery isn’t art is a rather curious reaction. Photography still seems to struggle with the perception that it isn’t art. I saw a comment not that long ago that a photograph wasn’t art because the photographer, unlike a painter or sculptor, didn’t actually create anything. Rather, the photographer simply recorded something that already existed. That commenter appears to have a point. If you Google the seven forms of art, photography is not consistently listed as one of the seven. (For some odd reason, cinema is occasionally included, though.)
Every definition of art I have seen includes the same three criteria: creativity, imagination, and skill. Fine art photography certainly embodies those three components. AI imagery wold also seem to incorporate those qualities as well. Someone has to use creativity and imagination to envision the image and then possess the skill to program the computer to produce that image. That sounds a lot like art to me, and surely there is plenty of room on the list of art forms to include AI imagery without displacing any of the others.
“Photography never lies: or rather it can lie as to the meaning of the thing…never to its existence.” – Roland Barthes
There have been opinions expressing that AI images are dishonest, that they don’t represent reality. The same thing has been said about photography. If you Google “honesty in photography,” you get approximately 573 million search results, so this is not a question that has sprung up only since the introduction of AI.
Photography has always been unfairly burdened with the flawed belief that every photograph should, and does, accurately and faithfully represent what we saw with our eyes. When that shutter button is pressed, the resulting image is exactly as it was at that specific moment in time.
The truth of the matter, whether one wants to believe it or not, is that no photograph ever captures exactly what we see. Every decision we make in setting up the shot — focal length, shutter speed, aperture, camera position, even what we decide to include or exclude from the frame — distorts the final image from the original scene before we ever get to the digital darkroom of our computers. If accuracy to reality is the barometer by which honesty is measured, then pretty much every form of art, including photography, could be considered dishonest, and I certainly don’t believe that.
Photographs are used in a number of ways, and the context of its use is an extremely important factor to consider when deciding if an image is “honest” or “authentic.” Some photographs are documentary or photojournalistic in nature and meant to record an event whereas a fine art photograph is often meant to convey beauty, feeling, or an artist’s feelings. These are not interchangeable.
I always enjoyed the photograph of the U.S. Marines raising the flag on Iwo Jima. Until, that is, I found out it was staged. Yes, the flag was raised, but the photograph was a re-enactment of the actual event. The actual raising occurred prior to the photograph. When the photographer saw it, he thought it would make a great image. So, he had them do it again so he could capture it. When I learned that, the image no longer had the same meaning to me.
Contrast that with fine art photography. Fine art photography, having art in its name, is, or should be, free of the limitations we place on documentary photography. It is perfectly acceptable for photographers to edit their photos in any manner they choose to represent the vision and message they are trying to convey. We, as viewers, may not like their message or care for their vision, but that does not mean the photographer’s artistic freedom should be limited.
Some comments have been more specific by claiming that AI images are being passed off as legitimate photographs taken on location. In those instances, though, it’s not the image that is dishonest; it’s the person making the false claim who is being dishonest.
Another area that falls within the bounds of dishonesty is the means by which AI software is being trained. AI software does not simply imagine an image and then create it like humans do. Instead, humans program the software to create an image within certain parameters, and then the software generates the image based on what it has been taught. But, how has it been taught? By analyzing millions and millions of other photographs. And, where does it find all of these photographs? From the billions (trillions?) of photos uploaded to the internet.
The issue with that process is that it completely ignores the original photographer’s copyright in the original work. David Holz, the founder and CEO of the research lab Midjourney, admitted in an interview with Forbes that the company has been using millions of images scraped from the internet to train its AI software without the consent of or payment to the original photographer. The company is now being sued. It was recently reported that customers who use Adobe’s cloud sharing and backup services were automatically opted in to providing access to their images for training Adobe’s AI software, again without consent or payment. While Adobe has since clarified that images are not being used for generative AI purposes, the issue remains that Adobe is being less than transparent in obtaining customers’ approval to use their images.
The dishonesty is not in the images AI produces. The dishonesty lies with the companies that are ignoring copyright or using the finest of fine print to take away an artist’s rights. So, don’t condemn AI images as dishonest; instead, denounce those who, all for the sake of money, make false claims about the images or who steal our photographs to train the software.
AI images are here, and they are here to stay. The business of photography will no doubt see significant changes brought about by the new technology. If photographers continue to focus on the human component with which AI cannot duplicate, new opportunities will most certainly open up.
Lost in all the discussion about the future of photography, whether AI images are or are not art, whether AI images are dishonest, the greatest threat to photography is being missed: the treatment of photographs and, by extension, photography as nothing more than a commodity meant to be exploited by others for their own gain to photographers’ detriment. That is the real and continuing threat to photography.