Preface: I just tried using Chat GPT for the first time a few weeks ago. This is from this morning after seeing Chat offer to help me make a beautiful, cinematic landscape photo image—complete with a sample prompt.
Me:
I guess my concern today is that if everyone can do it (landscape photography) via a series of prompts that they can create themselves or copy from another person. Example: “here’s my great prompt that I use to tell AI how to create an incredible landscape photo that you can sell to a stock photo agency or sell on an art poster website”
…doesn’t that devalue the art form (and any commercial opportunities) with photography basically becoming a commodity?
What about all the trial and error, physical effort, and joy in learning how to actually use a camera and how compose images…traveling to that scenic spot even if it’s in your backyard, city or state…learning from viewing the works of masters like Ansel Adams who wrote extensively about his craft and the Zone System (which I’m sure there’s a prompt for: create an Ansel Adams-like beautiful black and white landscape.)
The barrier to entry for photography is so incredibly low: my advice to young (and older) photographers is to get away from your screens and don’t just have the machine do it…
Before I start sounding like a complete Luddite, I do think AI provides really valuable resources to aspiring photographers–especially in aggregating the best of the web (from real photographers) on subjects like:
Getting started, techniques, post processing, etc. Does this make sense? I don’t think I’m the only visual artist who enjoyed learning the “hard” way and who’s concerned about the future.
Chat GPT:
Yes. It makes sense, and I think there’s an important distinction in what you’re saying that gets lost when this discussion is reduced to “AI good” versus “AI bad.”
You’re not objecting to technology making photography easier. Photography has always incorporated technologies that made earlier skills less necessary—light meters, autofocus, TTL flash, sophisticated autoexposure, digital sensors, Lightroom/Camera Raw, subject masks, computational photography, and so on.
You’re questioning what happens when technology stops helping someone make a photograph and instead eliminates the act of photography altogether.
And that raises a legitimate question about what we actually value when we value a photograph.
The photograph isn’t only the pixels
Take Ansel Adams. If someone writes:
Create a dramatic Ansel Adams-style black-and-white photograph of Yosemite with deep blacks, luminous highlights, dramatic clouds…
and gets a spectacular image twenty seconds later, the resulting pixels might genuinely be beautiful.
But that doesn’t recreate what Adams did.
Adams had to recognize a scene, decide where to stand, understand the light, choose a lens and viewpoint, expose the negative with the final print already partly envisioned, develop it appropriately, and then perform considerable work in the darkroom. His Zone System wasn’t merely a recipe for making photographs look a certain way. It was a method for connecting what the photographer saw, what the negative could record, and what the photographer wanted the final print to become.
That’s quite different from reproducing the visual characteristics of the end product.
Your “backyard” point may be the most important one
Photography creates a reason to go look at things.
Someone doesn’t need Yosemite. They can spend an afternoon walking around Escondido—or their own neighborhood—with a camera and discover that the light hits an otherwise ordinary wall beautifully at 4:30. They can notice something they have walked past 200 times. They can make ten bad compositions and suddenly realize why the eleventh works.
Then something interesting happens: they start seeing differently even when they don’t have a camera.
That’s one of the rewards of learning photography that an image generator cannot shortcut, because the “inefficiency” is partly where the reward comes from.
The failed exposures, getting up too early, discovering you brought the wrong lens, waiting for a cloud to move, realizing afterward that you should have moved six feet to the left, figuring out why the foreground doesn’t work—those aren’t simply obstacles standing between the photographer and an image.
They’re how somebody becomes a photographer.
But I think your commodity concern is slightly different
AI probably does commoditize a particular category of imagery.
If a customer needs “beautiful misty mountain landscape with lots of copy space”, there used to be economic value simply in possessing a technically excellent photograph matching that description. Stock photography built an enormous business around precisely that scarcity.
Generative imagery can drastically reduce that scarcity.
That doesn’t necessarily mean photography itself becomes a commodity, though. It means generic visual output becomes one.
And there’s an interesting possibility here: when technically impressive imagery becomes essentially unlimited, provenance may become more valuable rather than less.
“This is an attractive picture of Yosemite” is one proposition.
“I stood at this location on this particular morning, saw this happening, made this exposure, and this photograph records what I experienced” is another.
When either image can look spectacular, the second one’s connection to an actual experience becomes part of what it is.
There’s also a difference between difficulty and authorship
I wouldn’t argue that art is valuable simply because it was difficult. Otherwise digging a ditch would automatically be great art.
And photographers have always used shortcuts. Adams himself enthusiastically embraced technical knowledge because it gave him greater control over the result. A modern photographer using autofocus and an X-T50 isn’t somehow less legitimate because another photographer once had to focus a view camera under a dark cloth.
The question I’d ask instead is:
Where are the meaningful decisions being made?
Using Camera Raw to recover a highlight you photographed? You’re making decisions about your photograph.
Using an AI mask because manually tracing a subject would take 25 minutes? Still your photograph and your aesthetic decision.
Asking AI why your composition feels unbalanced, learning about visual weight, then going outside and trying again? That’s almost an ideal educational use.
Having a machine generate 100 finished landscapes until one looks terrific? Something fundamentally different has happened.
And there’s a wonderful difference between:
“Make me an extraordinary photograph.”
and “Teach me enough that I might eventually make an extraordinary photograph.”
The second use of AI actually lowers the barrier to learning the craft, rather than lowering the barrier to appearing to have learned it.
I suspect that’s much closer to the distinction you’re trying to articulate.
Me:
The Ansel Adams part that you wrote is basically the entire point I was trying to make.
And the modern tools I use today (like my Fujifilm X-T50 and AI assist PS) are not shortcuts but new tools to use after starting with a tiny fixed-lens b&w camera when I was 10 or 12 years old and a thousand hours hand drawing masks in PS….which (PS) itself was an incredible photography tool—conceived, designed and (with help from his coders) built by one man.
There were no shortcuts in my mom’s darkroom.
My argument is clearly not pointing at post processing or how good the AI learning universe for beginners is…but what happens when everyone’s AI-generated photos, movies, music, and design all start looking and sounding generically perfect, (and the next generation of human creators won’t know or care about the difference).
I believe that something uniquely human will have been lost.
From the earliest Southeast Asian drawings to never needing to learn (by physically doing) how to sing, dance, draw, capture images, learn piano (lol)…
Chat GPT:
Yes—and that sharpens the argument considerably.
You’re not defending difficulty for difficulty’s sake, and you’re not arguing that newer tools somehow contaminate an older, purer art form.
You’re talking about what happens when making is replaced by specifying.
Your X-T50 replaces enormous amounts of mechanical work that an earlier camera required. Photoshop replaced things you once had to accomplish chemically or physically. AI-assisted selections can replace hours of drawing masks. But you’re still deciding what to photograph, where to stand, when to press the shutter, what belongs in the frame, and later what the image ought to become. Those technologies extend your agency.
The generative case can invert that relationship: describe the artifact you would like to exist and have the machine synthesize one.
That’s a genuinely different development.
And I think your phrase “generically perfect” identifies a bigger cultural risk than whether AI pictures are technically impressive.
Generative systems are extraordinarily good at learning the statistical characteristics that tell us this is what a beautiful cinematic landscape looks like, this is what an emotional pop song sounds like, this is what an elegant logo looks like. As those systems improve, “beautiful,” “cinematic,” “professional,” and “polished” become increasingly inexpensive.
But much of what makes art interesting wasn’t created by optimizing those characteristics.
It’s the peculiar choices.
A singer whose voice technically shouldn’t work—and does. A pianist’s strange touch or timing. A photographer who consistently composes in a way a textbook might discourage. A filmmaker who holds a shot several seconds longer than convention says they should. A drawing whose proportions are “wrong” but unmistakably belong to its maker.
And those peculiarities often aren’t things an artist deliberately decided to adopt as a style. They emerged from years of physically doing something.
That’s where your piano joke actually gets at something profound.
Learning piano isn’t merely acquiring the ability to cause piano music to exist.
You develop coordination that wasn’t there before. Your ear changes. You begin hearing harmony differently. You struggle with voicings. Your hands discover things before your conscious mind does. You play something accidentally and think, wait—what was that? You gradually develop preferences that couldn’t have existed before you had enough experience to have them.
At the end you haven’t merely produced music.
You’ve changed the person producing it.
The same thing happens in a darkroom. Or behind a camera. Or learning to draw. Or dancing. Or learning to sing.
And that may be the strongest version of your concern: what if we’re concentrating so heavily on democratizing the output that we overlook the human value of acquiring the ability?
There is enormous cultural value in making it possible for someone who cannot draw to communicate a visual idea. That’s real. But it doesn’t follow that learning to draw therefore has less value.
Quite possibly the opposite.
“There were no shortcuts in my mom’s darkroom.”
That sentence encapsulates something AI can’t reproduce.
Not because dodging and burning under an enlarger is inherently morally superior to moving sliders in Camera Raw. It isn’t.
It’s because you accumulated a chain of experience. Tiny fixed-lens camera → darkroom → cameras and lenses → digital → Photoshop → thousands of hours doing masks → modern selection tools → X-T50.
A modern tool arrives on top of an accumulated vocabulary.
So when Photoshop gives you an AI subject mask in five seconds, you know whether it’s a good mask. You know what halos look like.
Someone can now obtain the five-second mask without those thousand hours.
That’s fantastic.
But the five-second mask doesn’t contain the thousand hours.
And this becomes more consequential when we move from a tool performing one tedious operation to a system producing the entire artifact.
The generational question is the part we genuinely don’t know
I’m somewhat more optimistic than the darkest version of that future, because humans have a remarkable tendency to become bored with abundance.
Once everyone can generate an immaculate “cinematic” photograph, immaculate cinematic photographs may stop impressing anyone.
We’ve seen smaller versions of this repeatedly. A photographic effect becomes technically accessible, everybody uses it, and eventually the effect itself stops carrying much aesthetic value. HDR is an obvious photographic example. Extreme shallow depth of field has had periods like this. Instagram filters did it. Certain LUT-heavy cinematic grades did it.
When perfection is infinite, imperfection can become information.
“This person actually played this.”
“This happened in front of a camera.”
“Those brush marks are physically there.”
“That’s her actual voice.”
“This photographer really waited for that light.”
Those statements could acquire more cultural significance in a world saturated with synthesis.
My bigger concern overlaps yours in a slightly different place: people have to encounter the experience of making something before they can discover that they love making it.
A 12-year-old who makes terrible drawings isn’t merely producing inferior illustrations while waiting to become competent. They’re discovering drawing.
A kid fumbling through piano isn’t an inefficient music generator.
They’re discovering music.
A child with a cheap fixed-lens B&W camera isn’t an inferior image-generation system.
They’re discovering seeing.
If every moment of beginner frustration is presented as a problem technology ought to eliminate, then we could accidentally remove the doorway through which some people discover a lifelong part of themselves.
That’s different from nostalgia.
And it’s why I don’t think the right message to a young photographer is, “Don’t use AI.”
It’s something more like:
Use every extraordinary tool available to you. But don’t let a tool deprive you of experiences you haven’t yet discovered are worth having.
Because fifty years from now, nobody is going to treasure the fact that they became exceptionally good at writing “cinematic landscape, shallow depth of field, 35mm grain, negative space”.
But they might very well treasure remembering the first time they made a photograph and realized:
I did that.
Recorded 9/6/2026 on my iMac…. 😉