In this episode of the Fine Art Photography Podcast: Addressing the elephant in the room — A.I. You’re probably already using it. But what do you think of it?
Full Episode Transcript:
In this episode — addressing the elephant in the room — Artifical Intelligence or A.I. You’re probably already using it. But what do you think of it?
Hey everybody. Welcome back to another episode of the Fine Art Photography Podcast. As I’m sure you already know, AI is rapidly infiltrating many aspects of our lives — and that includes photo software and even cameras themselves. How do you feel about that?
Is it still a real photograph if you use sky replacement and other AI techniques?
I’m not here to pass judgment. I just want to be part of the conversation. Let me say right up front, as the photographer slash artist, you have the right to manipulate your photograph any way you want. You can add or remove elements. You can alter coloration. You can combine multiple images. It’s all fair game. Unless you’re a photojournalist — where ethics require images not be manipulated, then as far as I’m concerned — do what you want with your own images. This is the fine art photography podcast — and art can be whatever the artist makes.
A lot of people seem to think that AI manipulation means the photograph is no longer the truth. But what photograph is really the truth? Ansel Adams made significant manipulations in the darkroom — Have you ever seen a straight print of Moonrise Hernandez? As an art photographer, I’m more interested in presenting my interpretation of a scene rather than the actual documentary truth.
Even so, I can’t imagine using an app to repaint my skies, or change it from midday to golden light, or many of the other things AI can do.
But there’s another side to AI.
Have you ever taken a smart phone photo that looked better than the same image made with your DSLR? That’s frustrating isn’t it? It’s happened to me more times than I can count. That’s because phones use computational photography — powered in large part by AI.
I use AI everytime I use Adobe’s content-aware fill to remove sensor dust spots from skies in my landscape photographs. Thanks Sony — your sensor cleaning tech is a failure.
AI is a great time saver on mundane tasks — things that I am happy to hand-off to the computer, which not only saves time, but also usually does it better than I could do it manually. Editing imperfections like dust spots in a gradient sky can be really dicey when done manually!
There are applications out there that use crowd-sourced AI to make clean enlargements, remove noise, and sharpen images.
Red eye removal is an AI function.
Portrait mode is an AI function that relies on machine learning.
Cameras are selling routinely now with software that can recognize human eyes, animal eyes, bird eyes, and all kinds of vehicles for motor sports photography.
Adobe Sensei AI can identify objects in photographs to add keywords to your metadata — that’s a real timesaver. It can also help with skin smoothing and things like that.
DXO DeepPrime software uses neural networks and AI to make precise corrections to RAW files — automatically correcting things like lens imperfections, chromatic aberration, etc.
But where is the line? Is it OK to replace a sky? Is it OK to change the color cast of your landscape from midday light to golden hour? Is it OK to add artificial bokeh?
What’s your opinion?
There’s already software in the pipeline that can automatically remove unwanted shadows, perhaps even change compositions. Phone cameras are leading the way with AI innovations, because the tiny form factor and extreme engineering involved makes software development a necessity. Sensors and lenses can only be pshed so far on a slim phone, but software can always improve.
Skylum – a leading AI software company and the creators of Luminar AI — argues that what it offers are productivity enhancements. These are all things that people have been doing in Photoshop for years — only now they can be done quickly and more efficiently. They are particularly suggesting that pros, who need to create images for clients, can save time and make more money by using the quick editing software. I can’t argue with that.
I’ve never used Luminar, but I’ve seen video demos and in the hands of experienced users, it really can provide very fast edits.
The day is quickly coming when software will be more important to photography than cameras and lenses — maybe even more important than the photographer.
The only bright side I can see as a photographer who sells prints, is that perhaps collectors will continue to desire authenticity. Maybe the reality of capturing a scene and presenting it authentically will make a difference. It may even push more people toward analog photography — where the negative is a tangible thing. But who knows. There are people paying millions of dollars for NFT JPEGs, so the whims of art collectors can be very unpredictable.
One of the most expensive photographs ever sold is Rhein II by German photographer Andreas Gursky. Not to be judgmental, but there’s not much to it. It’s a very large print, sold in 2011, and one of the reasons it’s important is precisely because it was digitally manipulated.
Well that’s all I’ve got for this episode. Thanks for listening. I’ll talk to you again real soon.
Thanks for reading