With the introduction of artificial intelligence, photography is at the edge of another precipice. The art form has been here before — with the invention of Photoshop, the advent of the digital camera, and the age of NFTs, to name a few. The question is, is the precipice of AI-based photo editing and image generation the edge of a cliff down which a fall would be detrimental? Or, is it the edge of a stairstep leading up to growth, if navigated correctly? To answer that, there is a bigger question at play: at what point is a photograph no longer a photograph but digital art?
Artificial intelligence in photography is a spectrum rather than a niche with easily definable solid lines. At one end of the spectrum is AI-based editing that simply saves the photographer time but accomplishes something the photographer would have done without the assistance of AI. For example, consider Lightroom’s select subject tools. The ability to select just the teeth or skin in a simple tap, for example, has saved me many hours of portrait retouching. Think of tools like AI-based sky replacement, select subject and similar tools.
At this conservative side of the spectrum sits many AI naysayers who swear they would never use it yet really love the eye AF on their camera, that is, spoiler-alert, AI-based. While it’s easy to denounce AI photography made 100 percent by DALL-E, it’s harder to criticize the tools that do nothing except save photographers’ time. Many of the latest time-saving tools from Lightroom, Photoshop, Capture One, and Luminar are AI-based. These tools, however, are notably different because they are 100 percent technical and do not interfere with the artist’s creative vision.
At the opposite side of the spectrum sits AI image generation, graphics that are 100 percent generated by a computer with the help mod a few text-based prompts. These images are much more challenging to categorize as photographs. Images created 100 percent by a computer and zero percent by a camera are digital art, not photographs. Some will even have trouble with these images being called digital “art.” But, among the photography community, the general consensus seems to be that images generated entirely by a computer are not photographs.
The questions for the photography community to answer lie in the middle of that spectrum, the gray area on images that are shot with a camera, but finished on a computer. At what point does a AI photography become digital artwork instead? When 25 percent is comprised of AI? 50 percent? 75 percent?
This is a question for the photography community as a whole to answer — not just one person. I’m just one of the many who should be having this conversation. But, I will add my voice to the growing symphony of photographers discussing where AI fits inside modern photography.
In the gray area of AI, determining when a photograph becomes digital art instead revolves around three questions. The first is who created the vision for the image, the photographer or the AI? For example, I’ve seen many AI-aided photographs that fixed where the image fell just short of the photographer’s vision. For example, the photographer couldn’t back away fast enough, and the image is cropped too close, so the photographer used AI to fill in the edges. This is an image that I would still call a photograph. On the flip side, I’ve also seen the randomness that AI generates used to inspire some or all of a piece of art.
The second question to consider is — would the same image be possible to capture in-camera? The question is similar to that of artists that mix photographs with hand-drawn Photoshop work. For example, Brooke Shaden uses a camera and then Photoshop to visualize some of the deepest human emotions. I would still consider her a fine art photographer because her Photoshop work is turning real but intangible emotions into something visual — and everything else is done in-camera. This question could similarly be applied to artists that use AI. I admire artists that use Photoshop to create visual representations of something otherwise intangible. I’m annoyed by Photoshop artists that take boring photos in-camera, adding almost all of the creativity later inside Photoshop.
The third question in determining where an image falls on that spectrum from photography to digital art is this: Does human emotion remain? Arguing semantics on a powerful image that tugs at the heartstrings is difficult. For example, photographer Sophie Gammond recently used AI to make shelter dogs with cropped ears whole again. The resulting images are still photographs because of the answers to those three big questions: Gammond had the vision, the same image wouldn’t be possible to capture in camera, and the resulting images are still full of human emotion.
The questions about AI in photography need to be answered by the photography community as a whole. Where on the spectrum is the line between photography and digital art? Drawing the line too conservatively risks thwarting artists with unique visions and big emotions. Drawing the line too loosely risks diluting the definition of a photographer as someone who creates images, not someone who creates photos with a camera.
The final consideration is this — how much does the label “photographer” matter? Is a mixed media artist or digital artist any less an artist than a photographer? What is more important — transparency in how the image was created or labels?