Photography AI as art

[MattKing]

You mean they weren't real armadillos ?

 

[Pieter12]

You mean they weren't real armadillos ?
View attachment 390079

Depends on what you're smoking.

 

[MattKing]

One of the most fun concerts I ever attended was Commander Cody at the Commodore in Vancouver - probably about 5 years after that album was recorded.

 

[TomR55]

Don't sweat the small stuff. You already know what you have when you hold one of your negatives in your hand: the authentic article. There are many interesting things you can do to that negative afterwards, but the original is the original.

True, in so far as my own work is concerned. I was thinking about the myriad of applications, such as forensics, scientific research, and other applications where authenticity and some epistemic gravitas is essential.

 

[Ivo Stunga]

Sure. But until recently it took specialists to convincingly alter an image.

Yes, tech advancements make this easier. Does value come solely from ease of replication?

And core question remains: if a given AI-contributed work carries impact, meaning and feelings across - art or e-fart? And do these technicalities in art matter at all?

When looking at an oil painting, do you value it only because it's oil on canvas + artist time regardless the quality and what it makes you think and feel, or do you like it for completely different reasons regardless the method of madness? Like subject matter, composition and something else?

Ancient Egyptian sculptures are very suspiciously very symmetrical. Would you value them less if Archaeologists found an Ancient Egyptian "3D sculpture printing machine", or is the object on itself a subject of art, regardless of how it was made?

Can a digital photograph be art?
Because it's so much easier to make and manipulate!

 

[Ivo Stunga]

I guess these questions get recycled with the arrival of each new technology.

Scribes/Monks screamed at printing press;
Painters screamed at plate Photographers;
Plate Photographers screamed at 135/Film folk;
Film shooters screamed at Digital photography;
And now everybody still alive screams at AI...

And we will continue to do so until we've grown accustomed to the new demon and a new visual satan arrives... Just a human condition.


To me: additional "Cool" points if human made a given work. Extra credit and that's all. I'm not moved by Starry Nights and Mona Lisa's - so I just don't care by whatever method they've been made...

 

[koraks]

Scribes/Monks screamed at printing press;
Painters screamed at plate Photographers;
Plate Photographers screamed at 135/Film folk;
Film shooters screamed at Digital photography;

Perhaps some did/do, but I'm not sure whether the screaming we're hearing is representative of what 'the masses' think. Looking at the adoption rate of AI (and the other innovations you mention), the screaming is apparently done by either a relatively small majority, or by a majority that screams and at the same time embraces what they're yelling at.

 

[Jim Jones]

But surely all this digitisation, AI and can't believe any image you see, is making chemical photograph unique in the history of art.

Soon after the first Daguerreotype photographs were made, chemical photographers were seeking ways to create images far beyond what could be captured directly on film or plates. Clouds could easily be added to bare skies. Group photos of people who never met in real life could be composited. Retouching covered many flaws in the subjects. It is said that photographs never lie, but we should never blindly trust a competent photographer.

 

[Alan Edward Klein]

Something along the same lines, which points to the same predominantly technical interest on behalf of the person who asks it, which automatically bars any subsequent meaningful exchange about the content or intent of the image as such.

I disagree. Many people who ask this question may feel they are being defrauded.

 

[Alan Edward Klein]

There was a time...

Especially when a lot of people shot slides and projected them. What you shot is what you projected.

 

[Alan Edward Klein]

Yes, tech advancements make this easier. Does value come solely from ease of replication?

And core question remains: if a given AI-contributed work carries impact, meaning and feelings across - art or e-fart? And do these technicalities in art matter at all?

When looking at an oil painting, do you value it only because it's oil on canvas + artist time regardless the quality and what it makes you think and feel, or do you like it for completely different reasons regardless the method of madness? Like subject matter, composition and something else?

Ancient Egyptian sculptures are very suspiciously very symmetrical. Would you value them less if Archaeologists found an Ancient Egyptian "3D sculpture printing machine", or is the object on itself a subject of art, regardless of how it was made?

Can a digital photograph be art?
Because it's so much easier to make and manipulate!

But a photograph is capturing a moment in actual time. It's not a painting from someone's mind. If you can create a photo without leaving your bed or needing a camera, then the whole point of photography is gone. Basically, you're drawing with a computer.

 

[Pieter12]

But a photograph is capturing a moment in actual time. It's not a painting from someone's mind. If you can create a photo without leaving your bed or needing a camera, then the whole point of photography is gone. Basically, you're drawing with a computer.

AI does not create photos it makes photorealistic illustrations based on photos. It is a perfectly valid tool and it will take work from many working photographers, just as photography greatly replaced illustration in the past. The danger of AI images is they are becoming more and more difficult to distinguish from photographs and can be used to create images of situations that did not occur while viewers are lead to believe they did. As opposed to art that no one sees as a true representation of reality.

 

[MattKing]

Photographs don't "capture a moment in time".
They create a two dimensional (usually) facsimile of what might have been seen by an observer from a particular point of view, at a particular point of time.
They aid in recalling what might have been, but they certainly don't recreate it with exactitude.
They can be quite useful, in the scientific world, because they permit reasonably objective comparisons between two different but related states/things, but they are merely helpful aids if one is attempting to explain or understand what is actual and real.
They can, however, stand alone as a source of beauty and can inspire all sorts of human reactions like anger, lust, hunger, sadness, etcetera.
Much like everything else that can have both utilitarian and artistic uses.
There is a reason that the uses of photography are strictly constrained whenever issues of proof or explanation/education are involved.

Traditional photography and AI generated images do share one profound weakness - far too many people trust both uncritically. That is where AI problems arise from.

 

[cliveh]

Photographs don't "capture a moment in time".
They create a two dimensional (usually) facsimile of what might have been seen by an observer from a particular point of view, at a particular point of time.
They aid in recalling what might have been, but they certainly don't recreate it with exactitude.
They can be quite useful, in the scientific world, because they permit reasonably objective comparisons between two different but related states/things, but they are merely helpful aids if one is attempting to explain or understand what is actual and real.
They can, however, stand alone as a source of beauty and can inspire all sorts of human reactions like anger, lust, hunger, sadness, etcetera.
Much like everything else that can have both utilitarian and artistic uses.
There is a reason that the uses of photography are strictly constrained whenever issues of proof or explanation/education are involved.

Traditional photography and AI generated images do share one profound weakness - far too many people trust both uncritically. That is where AI problems arise from.

But in our imagination they capture a point in time.

 

[warden]

But in our imagination they capture a point in time.

I am placing my trust in photography to capture moments in time for myself and my loved ones, so I sure hope it works.

 

[Vaughn]

But in our imagination they capture a point in time.

Aye, photographs record time to feed our imagination, but we exist at the point of time, a point of no duration with everything else being past or future.

Edit -- Perhaps we can capture a line of time, as no shutter is fast enough to capture a point.

 

[Alan Edward Klein]

Claiming that AI computer-generated picture is the same as a camera taking a photograph just does not make a lot of sense. Especially for us. We're photographers not programmers.

 

[koraks]

Claiming that AI computer-generated picture is the same as a camera taking a photograph just does not make a lot of sense.

Nobody did.

We're photographers not programmers.

Most AI artists aren't programmers. Most violinists aren't violin builders. It's the same principle.

 

[snusmumriken]

Talking to my son about this yesterday evening. He’s a third year undergrad in electronic engineering.

Apparently a lot of technical know-how, craft, understanding of how language works, and imagination go into creating the requests to the AI engine that generate the best AI images, or surreal videos like those linked in post #342. Whether that is more or less ‘art’ than photography has been hitherto, is open to (pointless) debate.

When I asked how the self-taught AI engine then uses the input parameters to draw on the source material and generate its graphic output, my son grinned broadly, his eyes widened, and he said “Magic!” Fundamentally the process is just matrix algebra (like the tools already in Photoshop etc), but to comprehend all the connections the AI engine has learned is too complex - if we could understand, we could replace it with a simpler, non-AI model. [Personally, I would think that’s a likely progression for smaller tasks, rather than building huge processors into portable gadgets.]

Less awesome are two characteristics of AI that result from its pixel-based approach: a lack of attention to the borders of the image (because there’s nothing outside the frame to relate those pixels to); and the weird morphing of one shape into another, which just happens as it grades one set of pixels into another.

All that is a lot more than I knew yesterday morning!

 

[koraks]

Fundamentally the process is just matrix algebra

This is a fundamental misunderstanding/misgiving. It's NOT matrix algebra. It's a neural network approach. It is correct that we can't understand the inner workings of the neural network any better than we can understand how a proper biological brain works - which means we can in principle understand it, but the degree of complexity stands in the way of a practically feasible approach of understanding it.
The key difference between matrix algebra and a neural network is that the former is deterministic and pre-determined; the latter by contrast exhibits emergent behavior. If you look at it from a distance and treat both a matrix algebraic algorithm (let's say something like a Markov chain) and a neural network as a black box, they might seem somewhat similar in that they can exhibit similar behavior. However, the inner workings, the way they're 'built'/designed and their capabilities are fundamentally different.
What this misunderstanding illustrates is how deeply rooted the misconceptions surrounding 'AI' are (understandably!) and how difficult this makes it for the present population to intuitively grasp the capabilities of this form of AI. I expect that this will change over time just like the present generations have managed to come to grips with digital technology, the internet etc - all things that the vast majority of people don't really understand thoroughly either, but most of us are fairly well aware of the possibilities and impossibilities associated with these technologies. In a similar vein, I think we'll also come to grips with practical applications of neural networks, and in the near future, quantum computing.

 

[TomR55]

This is a fundamental misunderstanding/misgiving. It's NOT matrix algebra. It's a neural network approach. It is correct that we can't understand the inner workings of the neural network any better than we can understand how a proper biological brain works - which means we can in principle understand it, but the degree of complexity stands in the way of a practically feasible approach of understanding it.
The key difference between matrix algebra and a neural network is that the former is deterministic and pre-determined; the latter by contrast exhibits emergent behavior. If you look at it from a distance and treat both a matrix algebraic algorithm (let's say something like a Markov chain) and a neural network as a black box, they might seem somewhat similar in that they can exhibit similar behavior. However, the inner workings, the way they're 'built'/designed and their capabilities are fundamentally different.
What this misunderstanding illustrates is how deeply rooted the misconceptions surrounding 'AI' are (understandably!) and how difficult this makes it for the present population to intuitively grasp the capabilities of this form of AI. I expect that this will change over time just like the present generations have managed to come to grips with digital technology, the internet etc - all things that the vast majority of people don't really understand thoroughly either, but most of us are fairly well aware of the possibilities and impossibilities associated with these technologies. In a similar vein, I think we'll also come to grips with practical applications of neural networks, and in the near future, quantum computing.

A very well written and informative response—one that demonstrates a more than “pedestrian” understanding of the technology. I think a little about these things (but being retired think more about making photographs …), and I encounter some recurring questions/concerns. Here’s one: Assume that the neural networks employed by current AI are functionally equivalent to that of human beings. Human beings differ from software in that we exist in a physical world in which living and no-longer-living are fundamental to our thoughts, behaviors, and everything that follows from these. Machines can simulate (or convince us by language) that they ‘understand’ the world, but “have no skin in the game.” Allowing such a machine (meaning algorithm) to make existential decisions seems problematic.

I think what really keeps me up a night is what I perceive as an uncritical acceptance of a technology that very few understand (to the extent that it is “understandable”) for the promise of unlimited profit to a few and control over the masses. And this with no plan for how we will adapt as a species.

 

[koraks]

Allowing such a machine (meaning algorithm) to make existential decisions seems problematic.

What you're saying reminds me of a concept used in psychology, philosophy psychiatry and (I suppose) neurology: "theory of mind". Simply put, AI as we have it now doesn't have a theory of mind. It walks like a duck, it quacks like a duck, but it doesn't feel like a duck. Since as human bystanders we can only directly witness the walking and the talking, we may be led to believe the 'feeling' is also there, leading to the conclusion that it really is a duck. In reality, artificial intelligence is very artificial indeed in the sense that it lacks a number of vital aspects of real intelligence. Theory of mind and in a broader sense a lack of ability to self-reflect are key aspects of that set.

Now, we can argue whether constructs like theory of mind and the ability to reflect aren't also simply emergent behavioral patterns that can arise in an artificial system under certain conditions. Indeed, I think this is conceivable and it's probably a matter of time before we see the first signs of machines exhibiting 'feelings' that are indistinguishable to us from the feelings of, say, your pet cat. And not because the robot has such a realistic rubber skin and natural motion control - but because its silicon brain actually shows behavior that's difficult for us to distinguish from our own mental machinations.

I think what really keeps me up a night is what I perceive as an uncritical acceptance of a technology that very few understand (to the extent that it is “understandable”) for the promise of unlimited profit to a few and control over the masses. And this with no plan for how we will adapt as a species.

I wouldn't say this is an age-old question, but it's a question that's been with us for quite some time. I recall back when I was doing my PhD (not so awfully long ago btw), I had a colleague working in the same department who was doing research on the ethics of new technology as discussed in popular literature. One of the cases he researched was Jurassic Park. The genetic engineering technology put forth back then by Crichton brought about comparable ethical issues as AI. And we've seen (still see) somewhat similar issues related to e.g. nuclear technology, which also exhibits that interesting mix of a profound promise to humanity, potential for catastrophic failure and abuse, and widespread misunderstanding of the actual technology itself.

 

[TomR55]

Thank you for your quick reply. Interestingly, I thought about “Shadows of the Mind” written by Roger Penrose and Michael Crichton’s "Jurassic Park" right after sending that post. I’ll have to think a bit about introspection as a (subsumable) emergent behavior.

 

[Alan Edward Klein]

Nobody did.



Most AI artists aren't programmers. Most violinists aren't violin builders. It's the same principle.

Yes, people here did claim AI photographs are similar as camera-taking photos. They're not. And violin builders don't write music any more than most photographers build cameras. It's not the same principle. An AI computer creating an image is no more photography than my 3-year-old grandson is a photographer because he uses crayons to scratch out a picture of sorts. Neither is photography.

 

[koraks]

Yes, people here did claim AI photographs are similar as camera-taking photos.

'Similar to' is already different from 'same as'. It involves questions like 'similar in which ways', and conversely, questions about dissimilarities.

And violin builders don't write music any more than most photographers build cameras.

You're again being wildly inaccurate/loose/nonchalant in how you rephrase things you apparently think are the same, but really aren't. Writing music and performing are different things. The majority of performing violinists don't write music. The majority of violin builders don't write music. The majority of composers don't play the violin. Different things. If you bunch everything together on a big pile, of course you're left with dung. But that's not the fault of reality - it's the fault of the person who refuses to or cannot distinguish between conceptually distinct issues.

An AI computer creating an image is no more photography than my 3-year-old grandson is a photographer because he uses crayons to scratch out a picture of sorts.

Again, sloppy reasoning, bunching dissimilar stuff together resulting in meaningless tropes.

Dung.