How will AI affect "making" versus "taking" photo's?

[lmans]

AI = image representations ..... Photography = taking pictures. The difference is humanity.

 

[Shunnothe]

A fair number of academic texts have been generated around the philosophical nature and the cultural role of photography … . At the end of the day, people will ignore these kinds of distinctions and pay for what they need, in the case of commercial work, and what they like, in the case of “art.”

This is not to say that such questions are useless or meaningless. As several posts on this topic have shown, “commercial” photography will be immediately impacted closely followed by “art” (for want of a better term).

For those who make photographs as art, the challenge remains: how to persevere, how to keep making images when one is unsure that anyone cares.






For those who make photographs as art, the challenge remains: how to persevere, how to keep making images when one is unsure that anyone cares. Interestingly, tools like https://overchat.ai/ can offer fresh ways to explore creative ideas and connect with audiences by generating unexpected perspectives, helping artists break through creative blocks and find new inspiration in their work.

That’s a thoughtful reflection. I think you’re right — the philosophical debates around photography (or any art form) aren’t meaningless; they help us explore why we do what we do, even if most people don’t consciously consider these questions when consuming images.


For artists, the uncertainty of whether anyone cares has probably always been there, AI or not. Maybe the best we can do is keep creating for the sake of the work itself — for the process, for personal discovery — and trust that those who connect with it will find it. In a way, that’s what keeps art authentic despite shifting markets or technologies.

 

[Arthurwg]

Some people think AI will destroy humanity. Others think it already has.

 

[TomR55]

Some people think AI will destroy humanity. Others think it already has.

Or, more likely, AI will make it easier for us to destroy ourselves? Perhaps it’s a variation on Einstein’s quote:"The world will not be destroyed by those who do evil, but by those who watch them without doing anything." ~ Albert Einstein.

 

[Arthurwg]

Or, more likely, AI will make it easier for us to destroy ourselves? Perhaps it’s a variation on Einstein’s quote:"The world will not be destroyed by those who do evil, but by those who watch them without doing anything." ~ Albert Einstein.

I guess Einstein really was a genius.

 

[f/Alex]

In many ways, this seems like a form of AI.

The modern thing that we think of as "AI" is generally an extension of machine learning algorithms taken to the extreme. I'm sure there is a camera with a "Progam" mode that is based on machine learning, but I doubt it's an F6.

Most types of modern AI can be divided up into a few different categories:
Machine learning is essentially trying to brute force its way through any and every problem it is given, it functions on a series of iterations, throwing one thing at the wall until one of them works, choosing the one that worked, increasing the required effectiveness of the solution, and creating a pyramid of solutions until the best one is left. Back in the day (ok not *that* long ago haha) I worked on a science fair project (well I mostly was moral support) using this, way before the AI boom. We also never used to call this 'AI', but the last few years have changed that.

The other major type is the Large Language Model, which is the base for Generative AI, which is what I think most of us here are worried about (and yes, I'm incredibly worried about it). I don't understand LLMs very well at all, but from my understanding, they're even more computationally heavy. They *should* be limited to one or two fields, although we often try and make sort of general purpose LLMs (chat GPT, Dall-e, etc), and that is part of what leads them to get so many things wrong.

Back in the day, it was argued that photography could not be art because it was made by a "machine." It seems that the same argument is being made about computer generated art today.

I have nothing against digital art, I have great friends who are in that scene, but a distinction must be made between digitally created art by human hands, and LLM-generated art. The latter lacks the emotions that go into any great piece of art. I figure this is what makes art distinct. I wish we lived in a society that valued the labour that goes into a piece of art, or even a product, over the materials, but we do not.

then I created a print out of the spreadsheet. Was I an artist making art?

I would argue that there is art in the banality of even a spreadsheet. Maybe not art that I would enjoy looking at, but. Part of my insistence against automation is that suddenly the product that you receive never had the human touch in its assembly. I'm also totally a weirdo.

("why are the computer kids getting all the dough and we aren't getting any, waaaaahhhh")

Nah I got friends who are adjacent to that field, no one's getting money. Or jobs. My technology friends are just as questionably employed as my liberal arts self.

 

[Arthurwg]

At one time it was said that "the art is the idea." I think that's true, coupled with an aesthetic "sensibility." Can AI come up with that combination? Probably yes.

 

[TomR55]

At one time it was said that "the art is the idea." I think that's true, coupled with an aesthetic "sensibility." Can AI come up with that combination? Probably yes.

From a limited understanding of how current AI products work… . These technologies appropriate ideas from a large corpus of data and are driven by probability in creating a response—but having no extension in the world, nor any “affect,” the “idea” is just a component that will be used in the creation of the “response.”

I liked the term “aesthetic sensibility” in your post. Our (human) aesthetic ‘sensibility' depends upon being ‘in the world,’ i.e., the human experience of time, space, living and dying, … . From our experience of the world, we act with intentionality; we create (so-called artistic) images out of human needs, desires, … images with particular ‘intents.’ Can an algorithm have ‘intents?’ If the answer is yes, then these must be “algorithmic” as opposed to human intents.

Our challenge is to produce artistic output, that which is “authentic” to human experience, different from that of the algorithm. Now this is particularly difficult to do; our audience is driven by the surface appearances, unless they are inclined or primed to “expect” something different; presenting many acclaimed pieces of photography outside of a museum or prepared setting will likely illicit a different response from the viewer. In this setting, the AI program will likely generate images that are more likely to dazzle and astound the viewer—AI is probably good a producing “eye candy.” Part of the solution requires an educated public … something becoming more difficult as we compete with the unending torrent of algorithm-generated content across the spectrum.

AI can make interesting and breathtaking pictures “of” many things. A critical question remains: can any AI program make authentic pictures “about” an idea. A challenge for human beings (both producers and consumers of “art”) is can we tell the difference? And, of course, the immediate follow-up question: will the audience care?

 

[Alan Edward Klein]

AI doesn't have a heart, soul, intuition, moral foundation or emotions. It's basically repeating what people have created previously.

 

[Alan Edward Klein]

AI reminds me of the toy Mr. Potato Head? It comes with different sets of plastic ears eyes, noses and lips and what you do is you take a potato and stick your selection of eyes, ears, noses and lips onto the potato, to make a version of the potato head that you like.

https://www.amazon.com/Playskool-Mrs-Potato-Head-Suitcase/dp/B0050QJT6U

 

[koraks]

These technologies appropriate ideas from a large corpus of data and are driven by probability in creating a response—but having no extension in the world, nor any “affect,” the “idea” is just a component that will be used in the creation of the “response.”

In how current AI works, there's no concept of an idea, even. There's just data and patterns. A new pattern is generated and it's mathematically compared to existing patterns; if the fit is high enough, the product is deemed to be good. Note that the patterns I mention aren't ideas - they're just combinations of content elements without any meaning; in fact, the 'identification' of such patterns is basically a black-box behavior, so there's no sentience in the process anywhere. Very simplistically put, in the 'mind' of AI, ideas do not exist. It's mind doesn't exist, even. So to answer your question:

Can an algorithm have ‘intents?’

No, not in the human sense, and not at present.
There's a philosophical discussion underneath it that could argue otherwise, mostly based on the notion that we could perceive or model the human mind as an intricate set of algorithms. However, that's very thin ice IMO since it's going to be very, very difficult to separate e.g. neuroscience from tautological constructs we use to describe phenomena. An argument in this direction would easily become an academic exercise of chasing one's own tail, in a way.

Ultimately, I think the 'solution' will be an easier one - we'll probably decide, sort of collectively, that the intervention of a human hand in the artistic process will be the determinant for defining something as conceived (as opposed to generated) art. One reason is that we'll have to have a criterion that people will be able to use and apply. As it stands, criteria that rely on an understanding of how AI works don't answer to this, since pretty much nobody in society has a decent understanding of AI. Whether a human was the defining element in creating something, however, is a criterion that anyone can also intuitively understand (even though there will be a grey area, of course).

Btw, arguably, our collective lack of understanding of how AI works may change to an extent, in the same way we've learned to grapple (of sorts, to an extent) with e.g. 'digital'. At the same time, it's questionable if we really have built a collective understanding, or whether we've just come up with a set of symbols that happen to work well enough to be able to put the technology to use. I doubt that such symbols may work well for AI in the application to something as 'philosophically deep' as art, however.

Well, fun musings. In the end, we'll just have to wait & see, buckle up and enjoy the ride.

 

[Arthurwg]

AI doesn't have a heart, soul, intuition, moral foundation or emotions. It's basically repeating what people have created previously.

That may be how it is now, but in the future? I can imagine a system that can come up with something new, based on an analysis of how something new evolves, and how it is valued. Jackson Pollock was not a genius, but he came up with a logical solution for an artist who could not draw and lacked other obvious talent, and now we value those drips.

 

[TomR55]

In how current AI works, there's no concept of an idea, even. There's just data and patterns. A new pattern is generated and it's mathematically compared to existing patterns; if the fit is high enough, the product is deemed to be good. Note that the patterns I mention aren't ideas - they're just combinations of content elements without any meaning; in fact, the 'identification' of such patterns is basically a black-box behavior, so there's no sentience in the process anywhere. Very simplistically put, in the 'mind' of AI, ideas do not exist. It's mind doesn't exist, even. So to answer your question:

No, not in the human sense, and not at present.
There's a philosophical discussion underneath it that could argue otherwise, mostly based on the notion that we could perceive or model the human mind as an intricate set of algorithms. However, that's very thin ice IMO since it's going to be very, very difficult to separate e.g. neuroscience from tautological constructs we use to describe phenomena. An argument in this direction would easily become an academic exercise of chasing one's own tail, in a way.

Ultimately, I think the 'solution' will be an easier one - we'll probably decide, sort of collectively, that the intervention of a human hand in the artistic process will be the determinant for defining something as conceived (as opposed to generated) art. One reason is that we'll have to have a criterion that people will be able to use and apply. As it stands, criteria that rely on an understanding of how AI works don't answer to this, since pretty much nobody in society has a decent understanding of AI. Whether a human was the defining element in creating something, however, is a criterion that anyone can also intuitively understand (even though there will be a grey area, of course).

Btw, arguably, our collective lack of understanding of how AI works may change to an extent, in the same way we've learned to grapple (of sorts, to an extent) with e.g. 'digital'. At the same time, it's questionable if we really have built a collective understanding, or whether we've just come up with a set of symbols that happen to work well enough to be able to put the technology to use. I doubt that such symbols may work well for AI in the application to something as 'philosophically deep' as art, however.

Well, fun musings. In the end, we'll just have to wait & see, buckle up and enjoy the ride.

Thank you for the detailed reply … makes me think of Weizenbaum’s Eliza—on steroids, and with a global user community.

 

[MattKing]

This should complicate the discussion
AI and robotics fight fraud

 

[koraks]

Thank you for the detailed reply … makes me think of Weizenbaum’s Eliza—on steroids, and with a global user community.

Yes, from a functional viewpoint, that's actually a pretty good parallel!