Rob Adams a Painter's Blog painter's progress

July 1, 2024

Through the Looking Glass

Filed under: Art History,Artificial Intelligence,Uncategorized — Tags: — Rob Adams @ 6:26 pm

There is a type of dystopia where most the citizens are happy and believe their dystopia is an utopia. If it’s literature then the rebel gets subsumed or brainwashed, or in the sci-fi world a lone rebel fights back and brings the system crashing down. The result is presumably that the ignorant masses can now live in a miserable but newly aware state. So is the Ai revolution taking us to Big Brother? The press and pundits would love this to be true. Or do we already live in a dystopia and Ai will open our eyes to the fact?

I suspect the answer is both of the above. We love opposites, they are easy, black and white good and evil, binaries are beautiful. However we live in a world of ever changing greys where a bad act can have a good result and a good one can pave that path to hell.

Ai has already changed our appreciation of images. many now view any image with suspicion. This is maybe a good thing, illusory images have always been used to tell lies. To sell religious myth and to gild the reputations of the powerful. This is only one side of the mirror though. They have also been used to tell stories of beauty and delight, of poetry and wisdom. Most of the deluge of images produced by Ai have amost no story to tell. Some are interesting and striking but they are few and far between. After a while you recognise all the different types of image. The Robots, monsters, heroes and heroines are all cut from the same cloth.

On the other side of the coin there are those who are obsessed with what they imagine to be realism. They mean photographic of course, the goal is an image that is indistinguishable from a genuine photo. What you do after this goal is achieved is unclear. I have no interest in the fake photos… advertisers however are rubbing their hands at the thought of no model or photographic fees. Hollywood is also preparing to jump in. With Ai you will very soon be able to make films with dead actors. John Wayne may ride across the prairie again, Marylyn Monroe might star in a sci-fi flick. Their joy will be short lived though. If the present arc of advancement continues you won’t need a studio, actors, writers or any of the enormous crews. You will suggest a story, pick your actors and an hour or so later your movie will be ready to watch. If you think it will be terrible, think again, the models will be trained on all the greatest movies. The Ai ones may not be Oscar winners but they will be more watchable than most of the trash released in a year.

For now though I was happy to play with this new toy. I tried mixing styles and subjects… not so easy as they need to be balanced. I found different ways of getting what I wanted rather than what the Ai gave me. I also made videos of my discoveries which became strangely popular for basic screeen captures with me gabbing in the background. Here’s a few of the more deranged images I made.

It is very good at baroque madness. The hard bit is controlling the composition and getting things in places you wish. This was done with a hand drawn depth map.

This is an example of the more unstructured random output. For this I input a series of conflicting images. Baroque buildings and paintings mixed with fractals I’ll put a few of the images below and you can try to work out what influences they had!

I even remade one of my old illustrations… I got increased finish but less charm.

It is very good at the generic. So horror, sci-fi and of course Steampunk.

Then there are the Ai people, they don’t come in very many varieties. The girls are sexy and the guys are musclebound.

The above demonstrates a point in my earlier post. There is no personal style in the Ai world. One thing I wanted to manage was an equivalent of my pen and ink style. So far this has been impossible… I don’t know whether to feel pleased or frustrated! Below is as near as I got.

I think that is enough about Ai for now. I might return to it as changes are happening at breakneck speed.

June 25, 2024

Tweedle Dum and Tweedle Dee

Filed under: Artificial Intelligence,Uncategorized — Tags: — Rob Adams @ 8:40 am

When the various digital developments arrived in the past they were almost immediately put to work. We got Pixar and other wonderful entertainments. After Effects and other 2D software was there behind the scenes. So what will Ai bring us? I did a few comic style tests and worked out that with a minimal amount of hand refine you could produce the artwork of a 30 page comic in a couple of weeks. So where are they? As far as I know technology doesn’t seem to be making inroads into storyboarding though it seems perfect for the job.

Despite complaining about their images being sampled the picture libraries are packing their catalogues with rather poor Ai images… mostly knocked up in Midjourney by the look of it. What I am getting at is Ai is somewhat short of practical uses. I see a few examples in adverts, but where are the Ai illustrated picture books? I did a few samples, to see what could be done.

This took me about an hour, based on a pencil sketch I did first.

I hate to think how long this would have taken to do by traditional means! If I was to use such an image I might rework the face by hand to remove the generic look.

I’ll put in a couple of silly comic strip tests…

These were a few hours work. You can see why artists are running scared. The strange thing about the ease of production is the drop in value of the artwork. Not only financial but in social value. People dismiss an Ai image after a brief glance. The very same image put before the same people a decade ago would have elicited Ohhs and Ahhs. Worse, I had some of my hand done pen drawings dismissed as Ai and scorned. People very much don’t want to be fooled and are suspicious of any imagery. Perhaps they don’t wish to be seen to accidentaly like an artificial image as that would discredit their powers of discernment.

I did a slightly immoral (and very unscientific) test of this and presented a hand done image and an Ai image for reactions. Except I switched them. The hand done image was duly scorned! The sheer scale of Ai image production is staggering, there are many millions of images produced every day. Far more than there are eyes to see them. I don’t want to sound pessimistic but it is hard to see it all turning out well. Oddly I think actual physical paintings might benefit. I suspect people won’t want Ai images on their walls, indeed they may not want anything that might even be suspected of being Ai.

So what is Ai image making for? Perhaps nothing in particular. It is a Swiss Army knife you could apply it anywhere. The thing which consumes current artists is that you need no particular skills to use it. Anyone with a keyboard who can type a short sentence can make an image in any style. To make a good quality image takes more practice but not the years of study learning to draw well takes. I have mixed feelings but there is no point in crying over spilt milk. The weavers faced with the Spinning Jenny and then power-looms saw their years of experience made valueless and now hand weaving is for most a hobby. Image making will I suppose go the same way.

Hand made art has in any case been in decline for a very long time. First came photography which automated recording the world. Although photographers talk up their skills it’s still mostly pressing a button at the right time. This reduced hand made image making to representing things which didn’t exist to be photographed. Still a pretty large market and imagination was brought to the fore. Then cheap reproduction took another chunk out of the required artworks. At each of these points the numbers of people (relative to population) gaining the skills and earning their living by them reduced.

I might leave the fine art world out of this as it is essentially a cult producing tokens for trading and possession. Rather like the trade in religious relics, the item itself has no value, it borrows its value from its context: The rich person’s apartment or the gallery wall. However it might be considered another nail in the coffin of skilful art, the prestige of being a creator could be claimed by anyone with little effort.

I will end with a couple of random observations. Ai art seems to be producing no stars. Illustrators have a following, some draw their comic books on line so their fans can watch the process. Ai produces people who demonstrate the process. There is little interest in the final image only in the workflow which made it. Indeed the workflow might be considered the artwork rather than the image it creates. So there we are, a looking glass world. We are still in mid-air falling down the White Rabbit’s hole, what sort of landing is ahead is anybody’s guess.

Older Posts »

Powered by WordPress

error: Content is protected !!