Is photography dying? Part 2: evolution

[Tom R]

Consider two facets of a photograph: its production and its consumption. How an individual produces a photograph is certainly important to that individual producer, but has no impact on how it (that same photograph) will be consumed: In consumerist societies (such as the United States) the "value" of an object is transactional.

Decades of rampant consumerism has resulted in the systematic erosion of "value" as we (producers) define it.

Now, Generative AI challenges producers' notions of authorship, authenticity, etc. But again, by the transactional nature of any photograph as a potential product of consumption the reputation of the producer, which translates into potential sales/barter value the product, overshadows the subject or "aesthetics" of the image. The only possible exceptions to this withering process are photographs whose value is directly dependent on their accurate rendering of things in the world; I am talking here about photographs made for important documents, as well as medical or forensic images. Sadly, these images will most certainly be made by non-human actors in the very near future. Capitalism is a race to the bottom ... .

I am not sure where any of this leaves us. Speaking for my work: I enjoy what I do; it provides purpose and I keep the negatives.

 

[RichC]

In short, you as a photographer care whether a photo-realistic image is a real photo or an AI fiction, but most people won't - with a few exceptions such as passport photos...

 

[jsrockit]

In short, you as a photographer care whether a photo-realistic image is a real photo or an AI fiction, but most people won't - with a few exceptions such as passport photos...

Family photos? Vacation photos?

 

[Harry the K]

Family photos? Vacation photos?

When photography is roughly divided into "private use" and "commercial use", only the latter department is affected by the use of ai.

 

[raydm6]

There’s a very interesting thread on Photorio using a newly released Midjourney “/describe” command function.

The generated imagery in the Photorio post is interesting but viscerally disturbing/spooky. Can’t quite put my finger on it.

Quite compelling technology nonetheless.

AI image description with MIDJOURNEY

For a few days now, MIDJOURNEY has also had an image description function. However, there you are not dependent on well-known images with clear titles. You can also upload your own motives and have them described. The command to this is : >>describe [picture]>person on march22, 1961 the...

www.photrio.com

Midjourney Flips the Formula with New Image-to-Text Generator

Midjourney can now describe and remix images.

petapixel.com

 

[raydm6]

Here’s another one; AI video generated from your photographs:

Zain Kahn on LinkedIn: This isn't a drone video. It's not even a real video. This is footage… | 236 comments

This isn't a drone video. It's not even a real video. This is footage was created entirely from a series of photographs by an AI. It sounds unreal, but a… | 236 comments on LinkedIn

www.linkedin.com

 

[Pál_K]

Consider two facets of a photograph: its production and its consumption. How an individual produces a photograph is certainly important to that individual producer, but has no impact on how it (that same photograph) will be consumed: In consumerist societies (such as the United States) the "value" of an object is transactional.

Maybe to most people. Even as a consumer, I value how the photographer produced the photo, whether glass plate, Polaroid, film, or digital, as well as the image itself. I may not be able to discern 35mm or medium format from high MP digital, but when I recognize a Tri-X or Kodachrome image, I appreciate it.

Sadly, these images will most certainly be made by non-human actors in the very near future. Capitalism is a race to the bottom ... .

Capitalism is what gives us ever-improving products fostered by competition; Socialism “works“ on a small scale only as a parasite on a foundation of Capitalism. Competition and the rewards of a successful product are why Nikon, Canon, Leica, Pentax, Olympus, et Al., became progressively better from the 1950’s onward (not to mention things such as personal computers and cellphones). Yet, you’re correct in a way: these same advancements give us distracted people driving with intense high-beams on constantly, and families and friends sitting in a restaurant or coffee shop staring at their cellphones instead of normal social interaction. When AI can generate images such Ansel’s Moonrise and other works effortlessly and prolifically, maybe we won’t value such images at all.

 

[Henry]

I think of things like MidJourney as a sort of scratch pad for ideas. I think it’s interesting to see what it will do when fed prompts with different classic, well documented, rangefinder lenses. I won’t post them here, because I think it’s a bit iffy, but it gives you an idea of what the “convergent look” of a lot of classic lenses is.

 

[wlewisiii]

Never believed it will go away any more than painting did.

There will always be new stuff.

Nothing new for photograrphers in that.

 

[Henry]

Never believed it will go away any more than painting did.

There will always be new stuff.

Nothing new for photograrphers in that.

Oh absolutely. Digital didn’t kill photographers and neither will AI. I think we had the same debate when cellphone cameras became more machine learning photography simulators, but it turns out it’s all art. (At least, imo)

 

[dbell]

When AI can generate images such Ansel’s Moonrise and other works effortlessly and prolifically, maybe we won’t value such images at all.

But will it mean anything to anyone if we know that no human being ever stood there and witnessed that scene, even in their mind's eye? AI doesn't "create" anything. It just mashes together whatever it's been trained on and spits out whatever it's asked for. There is no "intent" or "imagination" involved (whatever those are...). Sure, I might be able to say "give me two different images that combine elements of Michelangelo and Van Gogh," and have it generate two files. What it won't be able to do is answer when I ask "which one is better?" It also won't tell me that the entire exercise is ridiculous to begin with...

If art means anything, and has any value at all, it's because it constitutes HUMAN expression. It's a way of responding to our experience of the world. Machines have no such need.

 

[Sonnar Brian]

I prefer using older lenses, ones designed without the aid of a computer. I'm getting an image formed through the optical design of a Human that had a vision of how they wanted the lens to render. AI is a program, written by a team of software engineers. They put their vision of rendering into the code.

 

[Deleted member 82967]

I don't understand the notion that a picture created by an AI algorithm is somehow the same as a photograph.

Hello neighbor! Me neither.

Noticeable in this thread is the majority usage of the term "photograph" over "image", "photography" over "imagery", etc. That means that this thread is applying some spin to the topic or sub-topic anyhoo.

 

[Dogman]

AI really doesn't create anything. It just spits out information. That's not "creating". That's reciting.

But in the current cultural climate of the world, I'm not sure that matters. It will be interesting to see if humans of the future will know the difference. Or, if they do, will they care.

 

[jsrockit]

AI really doesn't create anything. It just spits out information. That's not "creating". That's reciting.

That would mean making a collage isn´t making something or using sampling in music isn't making something. Using AI to make Art is just another form of Appropriation and Recontextualization. It is nothing new in the world of art. You use text to tell the system what to look for and it jumbles it all together from many different sources to spit out something new. If you like what it looks like, you use it. If you don't like the results, you tweak your text. I'm not saying I want to use this method to make images or that it is fun, but Artists are already using it to make things they could not with photography.

How Will AI Transform Photography?

As artists experiment with this fast-evolving technology, they uncover creative opportunity, absurdity, and bias.

aperture.org

 

[retinax]

Yeah ai might compete with commercial photography in fields where photos are used as illustration, i.e. not to depict a specific person, product or place. The vast majority of photos is taken to say "I was here", "I saw this", "kiddo really was this small", "kitty is sooo cute" or "look what a high status person I am". I don't see how AI can replace any of that, except where I is used to deceive. All these claims rest upon the notion that something real is depicted, however spurious the connection between photo and reality may be. Open or too widespread deceptive use of AI imagery would undermine these claims and make the use of any imagery, ai or photographic, useless for these purposes.

 

[Deleted member 82967]

It could be that "AI", like "exposure", "resolution" , "IQ", etc has more meanings than implied in this thread.

Is an image upsized by Topaz "AI" real or not?

Going just to ye goode olde JPEG is a JPEG "real" or not?

Is any RGB image converted by a computer from raw real?

Is Nikon's evaluative metering real?

 

[Dogman]

That would mean making a collage isn´t making something or using sampling in music isn't making something. Using AI to make Art is just another form of Appropriation and Recontextualization. It is nothing new in the world of art. You use text to tell the system what to look for and it jumbles it all together from many different sources to spit out something new. If you like what it looks like, you use it. If you don't like the results, you tweak your text. I'm not saying I want to use this method to make images or that it is fun, but Artists are already using it to make things they could not with photography.

How Will AI Transform Photography?

As artists experiment with this fast-evolving technology, they uncover creative opportunity, absurdity, and bias.

aperture.org

It's the human input that makes appropriation, collage, etc., into art. It involves human decision making at each step. Just telling a machine what you want and letting it make the decisions is pretty sterile, uninvolved factory work. When Duchamp wrote "R. Mutt" on a urinal and put it on display, it was his impish input and humor that made it art, not the urinal itself.

AI can be a powerful tool for humans but on its own it just a tool, not a new artist of the brave new world.

 

[jsrockit]

It's the human input that makes appropriation, collage, etc., into art. It involves human decision making at each step. Just telling a machine what you want and letting it make the decisions is pretty sterile, uninvolved factory work. When Duchamp wrote "R. Mutt" on a urinal and put it on display, it was his impish input and humor that made it art, not the urinal itself.

AI can be a powerful tool for humans but on its own it just a tool, not a new artist of the brave new world.

A human is inputing text into a machine to make new art. I never said the machine makes art without human instructions. You are telling the machine what you want. It is easy to use badly. It is easy to dismiss. However, it could output something new and radical based on what you input. I'm not talking about facsimiles of art we already have. The Fluxus art movement sometimes just used text on paper to say what an artwork would be without actually making the artwork. The text was the artwork. I see this as a continuation of that. Except you can see how a machine interprets your instructions.

 

[JoeV]

But will it mean anything to anyone if we know that no human being ever stood there and witnessed that scene, even in their mind's eye? AI doesn't "create" anything. It just mashes together whatever it's been trained on and spits out whatever it's asked for. There is no "intent" or "imagination" involved (whatever those are...). Sure, I might be able to say "give me two different images that combine elements of Michelangelo and Van Gogh," and have it generate two files. What it won't be able to do is answer when I ask "which one is better?" It also won't tell me that the entire exercise is ridiculous to begin with...

If art means anything, and has any value at all, it's because it constitutes HUMAN expression. It's a way of responding to our experience of the world. Machines have no such need.

I am reminded of things like surveillance cameras, constantly recording things on its own. There’s a curatorial opportunity here for real human creatives to cull select images or footage out from the detritus of machine-generated imagery.

Like when I attached a cardboard box in front of my CRT television on the night of the 2008 elections and recorded freeze-framed VCR images onto pinhole camera paper negatives. I didn’t create that endless broadcast stream of election coverage, but instead made something of my own from the raw feed. This may be how artists will live with AI in the future. (I would’ve posted that pinhole photo but can’t find it at the moment!)