With the capabilities of image-processing software increasing every day, what does photography have left to offer?
Today: sky replacement. No, wait: automatic sky replacement. No, wait again; sky replacement with lighting adjustment. “AI art” from DALL-E, so engaging that people say “wow” and photographers spend hours fiddling with it. “Filters” that make people look like celebrities with perfect skin. “Upsizing” software that spins out new detail to fill the gaps.
It’s only getting worse (or better, if that’s how you see it). Eventually, people will be able to get any image they want right off-the-shelf, then adjust it to taste via sliders with labels like “Colorful”, “Moody”, “Glamourous” or “Edgy”.
Artists will, of course, find ways to work with all this. But what’s left for photographers? Two things, as I see it: authenticity and originality.
Authenticity just means an actual photo. I’m not a “straight from the camera” guy – we can clone out distractions, combine exposures, adjust dynamic range, enhance colors, all sorts of things. But the end result has to feel real: it’s an actual place, it was really raining, it’s a genuine object aged by time, the bird was really there. Yes, software can create those illusions, although “deepfakes” of celebrities are easier than believable synthetic landscapes. Authenticity is about stuff that’s impossible to explain or quantify; we just know it when we see it.
Originality doesn’t mean something that’s never been done before; it just means the idea came out of your head. You decided what to shoot and how to get it, you did the composition, the processing. It sn’t a mashup of stuff scraped off the web by a guilt-free “AI”. You didn’t just click a button labelled “optimize”. It’s intentional, not the output of algorithms.
Having said all that (and you may not even agree) here’s a photo I did recently, which I will say is authentic and original. I did extensive curve adjustments, pushed areas to black, combined exposures (to bring in more lights on the highway), and adjusted color to convey how the city felt as I stood there in the dark, listening to the traffic, and maybe reflecting a bit on the 50 years I’ve lived here. Not everyone will like it… hey, that’s the best part.
Guilty as charged, I have been dabbling in some AI and having fun doing it. Sometimes I use my own images as a starting point, other times it’s something entirely out of my imagination. Honestly an AI piece can take me as long as one of my original photographs to get it just the way “I” like it. Hours even.
That being said, it can’t beat a beautiful original image like yours here. The lighting is exquisite! Impressive and I don’t think AI can create that, well at least not yet.
After fiddling with the Stable Diffusion software for a few weeks (after realizing I do NOT like DALL-E) I can say that getting an image that reflects what you want on the first batch is totally a matter of luck. There are a couple things I am still trying – and failing – to get an image for, and in all honesty it might be faster to just draw it myself. WHEN it works, the program-generated image is faster than doing it by hand, but the chance of getting what you want without any weirdness is not as high as the fans like to claim.
I can now say that with some authority, having tried it myself.
Katrina, you just gave me an idea for another post in the future.
What’s left (behind) is what we started with: a photon concentrator/bender/projector and the back wall of a mufti-lined box—together a producer of light graphs for an operator called a photographer.
What we honor now is elaborated process and artifacts thereof. Faith in the documentary (recording) function of photojournalism survived dodge and burn but succumbed to Pshop and the App Store.
Keep the faith, Jack. What was old will be new again. Tastes change.
Jim, these are good points. First and foremost, I like the image you presented. Second, I have felt lately that I am more attracted to simple images that look natural and “look” straight out the camera, then to “wow” images I used to be more attracted to, as a viewer. I have learned to embrace simplicity in my taste as an observer. I think images that reproduce what the eyes can see, whether adjustments were done or not to the initial photograph, sometimes they are needed because the camera still does not see what the eye sees, will always have a place in people’s hearts.
We’ll be absolutely flooded with “AI” images for a while. And then, people will get tired of that whole thing. It always works that way.