Is AI-generated "art" actually ART?

Ditto for me re: Hockney. He is actually quite brave and creative imo in his new direction and uses of technology. He is also old enough, and famous enough, to do what he likes now, and it will be richly rewarded. His infamous childlike ipad-made Piccadilly tube sign was part of Sadiq Khans “Lets do London” 7 million dollar campaign. (Not to critisize the art, but just imagine how many NIH hospital beds that could have payed for).

1 Like

Probably. Thing is, we humans like to look for meaning in things.

I suspect that any old tomfoolery piece of ‘art’ created entirely by AI can be reinterpreted, given meaning or depth simply by the viewer’s imagination. There are no limits to the latter.

We can show 10 people the same artwork and more than 10 aspects or messages will they attribute to it. The artwork itself may no longer be the important thing. What we make of it may be.

Rather makes the artist redundant if the message can be fabricated without him/her but he ho

:woman_artist:t2:

Now you’re talking. Artistic license!
:grin:

2 Likes

I’d argue that the Giverney umbrella is principally a cultural token used for public display unlike domestically displayed souvenis. I also think that it’s more likely to evoke Monet’s garden rather than the water lily paintings done therein?

Love your image Ken, great example of using the tech! Thanks for explaining how you use various photomanipulation, and how AI itself does not produce usable images. I also find the AI generated images strange, slick and cliché. Still in the “uncanny valley”.

One of the main things about AI, SO FAR, is that we humans have to code it, direct it, ask it questions, give it images/music/material to work with.

I wonder, though, is that because AI is still learning, is it still at the 3-year-old child level of “a square house with a triangle roof and two windows and a door” stage? Will its slick surfaces, seemingly gleaned from teen video-games, marvel and anime cartoons and fashion photography, eventually develop into different aesthetics/styles that are actually creative?

What happens when it starts asking US questions, when it becomes sentient, the supposed Singularity? What will its motives be, besides its own existence? Will it have an abstract, mathematical and logical approach to everything (eg keeping itself running, expanding, perhaps finding resources necessary, self-repair etc), or will it have an equally mathematical and machine abstract notion of BEAUTY, of rightness, or a frightening efficiency which will find us humans unworthy of attention at all?

2 Likes

Apologies for my previous reply - fading memory - it was sixteen or so years ago. The Weng Fen photos were mounted not printed on aluminium. Maybe I should explore further!

1 Like

I’m not sure that public display is a factor but cultural token, possibly, given that Tokyo loves their Monet lily paintings. (NMWA gift shop!)

I have such an umbrella, and whether others recognise it as anything to do with Monet matters to me not one jot. What I love is that a basically gloomy rain day is rendered serenely green in my little space beneath its beautiful wings. I smile every time it unfurls

1 Like

I would love this umbrella

But then again, I would almost certainly lose it
:disappointed_relieved:

1 Like

Had a brief play with Stable Diffusion while waiting for a breakfast egg to boil (as you do…)

Given Hockney’s been popular on SF recently, I entered, ‘Hockney Catholic Madonna’ (‘Catholic’ obviously an essential filter!)

SD then generated a set of four images that I’m still thinking and deconstructing for the following reasons:

Hockney Madonna 1 is very different to the other three.

This is the closest to what I expected, but it’s also far worse than I could have imagined! Don’t know where those animé eyes came from, as they seem to have nothing to do with any of my filters.

The most Hockneyesque aspect is the palette, but I was puzzled by the face which isn’t that typical of his portraits. However the answer lies in the first image and therefore most common example to come up on a Google image search which was a portrait of Barry Humphries used on the publicity for Hockney’s portrait expo at the Bilbao Gug.

For me, the most interesting aspects are in the AI reworking of the tiled floors of C15th paintings that mark the discovery of single point perspective. It reminds me of how 2D renderings of complex 3D wireframe models sometimes implode and create objects that are geometrically impossible.

The other three images are very different, and for me far more interesting, particularly the bizare central one (love the enigmatic golden figure on the right), The RH image may be the first depiction of a Cross on Madonna’s robe - wonder where that came from? Note also the inability to successfuly use perspective - in the LH image, the RH tree is in front of the Madonna.

Lastly all the religious source material is from the Late Gothic / early Renaissance perhaps because that painting was less spatially complex and thus closer to the comparative flatness of much of Hockneys oeuvre.

The mistakes and ineffective spatial interpretations of these renderings make them interesting because one wouldn’t normally encounter images as strange as these.

Ah Ha! Someone has been letting AI read The Guardian

1 Like

Excellent exercise. Thank you! It certainly makes us think.

I rather like the flat plane portraits. Reminds me of those China Trade portraits done in Chinese workshops in Canton.

Strange basic error regarding the tree overlapping the LH Madonna. And, why is only one, the RH Madonna suddenly sporting a giant cross, pre-crucifixion? Why omitted on the others? And the direct gaze? Spooky

Maybe AI mixed things up with Madonna Ciccone?

1 Like

The flat colour areas are also Hockney-esque. The female face on the chest of the middle one is disturbing, as if it was a t-shirt image, as is the vaguely small human robed figure at the side. Also, on the right, the weirdly armed child with the pink ear looking at the Madonna. Its almost as if it knows that a Madonna is a female robed figure with a headcovering, in some kind of formal frame-like architectural setting, and possibly with a child. Strange that there are no infants. Maybe try it backward and ask for an Greek icon Madonna, David Hockney? Every time someone uses these programs, and makes selections of which images to refine, the program learns more. We are teaching the fox how to get into the henhouse. Fascinating!

1 Like

Well noted - forgot to mention that! Prior to Manet’s Olympia convention dictated women should be portrayed ‘modestly’ with a slightly lowered gaze.

The SD Hockney Madonnas reminded me of Tom Phillips screenprint, After Raphael from 50 years ago.

Except of course this is far more coherent

I used Hotpot for this one; Madonna and child, Vincent Van Gogh. Rather literally, it made Vincent the child, and we get a it of background starry night. Also, too many fingers on the infant. This seems to happen often with AI, not sure why, with all the images generated. Hmmmm. Is it having us on?

I “enhanced the face” using the online tool. ; “Enhance and repair damaged faces with the latest technology. AI image generators and old photos often contain faces distorted in several ways: hazy colors, unnatural eyes, blurriness, scratches, and asymmetrical features. Our AI corrects these flaws in seconds, producing more natural images.” Actually, I’ve noticed the asymmetrical features on alot of AI too. And the eyes are all still strangely wrong. And still too many fingers.

Durer’s woodcuts and Rembrandt’s etchings were made some time ago!

Are you familiar with Walter Benjamin’s, inter-War essay, The Work of Art in the Age of Mechanical Reproduction in which he discusses the demise of the ‘aura’ of the unque art object (in the context of the ‘new’ mediums of photography and cinema.

benjamin.pdf (mit.edu)

Although his argument has several flaws and he’s arguing for the reverse set of values to yours, his concept of ‘aura’ is a useful one in terms of your and @AnAussie 's arguments and I often use ‘auratic’ in this context. It’s a handy little word!

2 Likes

Thanks! Imagine you have been set a challenge to produce a picture of a Madonna - as DrMarkH has shown, AI doesn’t always get it right. The images could be manipulated in order to create a new image but that would involve a lot of human intervention. In the RH picture I can’t think why the anatomically challenged boy is in the picture at all. I haven’t tried Stable Diffusion but with so many errors in the examples shown I am not in any hurry. I must add that the leading AI companies do create photo-realistic images and are considerably better than " the 3-year-old child level of “a square house with a triangle roof and two windows and a door” stage" but even so, they are prone to errors too - such as curved, additional or missing limbs, To illustrate what I mean is a picture of a ballerina in the style of William Blake - an unaltered AI monstrosity.

I’ll attach a much earlier piece of mine called “…and the cow jumped over the moon”. just to illustrate the photo-realistic images which are now possible (the cow was produced by AI)

Finally that spark of creativity, imagination and sheer inventiveness present in art and music for example is entirely human and that will be missing from AI for many years to come.

2 Likes

Thank you @DrMarkH for the MIT article link. I will read. And yes, I know what fun was to be had with printing in 15th-16th century. Young Durer, still in his twenties may be said to have been inspired by the new technology and it’s ability to disseminate art.

As for the other AI art productions, I am afraid to say…. No, just afraid. A bit like disembodied AI voice recordings, they approximate reality with alarming misses. Too many fingers is a frightening error. The programming has clearly missed many things but that wouldn’t explain the bizarre inclusions, like that T-shirt image Madonna for one.

My worry is not so much about AI producing artworks. We can still be the judge of those. My worry is by what process is it learning about what humans like and want to see. Will it then begin manufacturing an alternative reality? Or indeed, pseudo humans with more fingers? So handy

:clown_face:

2 Likes

Looks like you’ve been reading today’s NY Times

Opinion | A.I.: Actually Insipid Until It’s Actively Insidious - The New York Times (nytimes.com)

You might find Shiner’s The Invention of Art, useful, it’s an analysis of the historical process whereby art and craft became separated in Western culture and a hierarchical relationship between the two evolved. The current best value seems to be Amazon UK s/h paperback - I think you’d find it a useful historical analysis of the separation of art from craft in Western culture.

1 Like