Does my (a.i.)art look big in This?

[The Legendary Shark]

Last two, promise. I'm getting bored with this myself, to be honest. It takes too long to get an image that's suitable, let alone close - and never spot on. I ended up trying to change the "joke" to fit the image, which isn't what I really wanted.

So I think I can call this experiment a failure, especially in light of V14's insightful "scrappy amateurish scrawl" comment.

But AI must be good for something for an aphantasic artist like me, so next I'm going to try using it for something else.

Stay tuned!

[Jimmy Baker's Assistant]

But AI must be good for something for an aphantasic artist like me, so next I'm going to try using it for something else.

I don't think AI art is ever going to give you what you want. Unless you want 150 pictures of Power Girl with 6 fingers. Then you're massively in luck!

[M.I.K.]

I'm not saying aphantasia isn't a drawback for visual artists, ('cos it stands to reason it would be), but there's a surprising amount of really quite accomplished artists with the condition, including the bloke who designed Disney's Little Mermaid and Beast from Beauty and The Beast.

Inversely, there's a lot of folk with a very vivid mind's eye who can't draw for toffee, (or any other confectionery of their choice), in the same way that you can't just stick most folk down in front of a landscape with a bunch of paints and expect them to produce a masterpiece.

[Vector14]

But AI must be good for something for an aphantasic artist like me, so next I'm going to try using it for something else.

I don't think AI art is ever going to give you what you want. Unless you want 150 pictures of Power Girl with 6 fingers. Then you're massively in luck!

The worry for me, is that in the not so distant future AI art will have developed so it gives people exactly what they want.
And it will be impossible to distinguish from real art.

[The Legendary Shark]

I dunno. Maybe it will force real art back into the real world, onto paper, canvas, and clay - driving up the value. Digital art, I feel, is most at risk - but I also feel that digital artists will find a way to use AI (as does the speaker in M.I.K.'s "A.I. Art book Scam" post on the Implications thread). Artists are imaginative and will find ways to use this freed genie. Unfortunately, scammers, cheats, and general ne'er-do-wells are also imaginative. The world will adjust, as it always has.

The artists known as scribes who copied out volume after volume may have been eventually rendered obsolete by the printing press, but it also ushered in new artists like wood-carvers for illustrated volumes and later cartoonists and photographers. That process took centuries, but this A.I. thing won't take long at all, I expect. From now until the dust settles might take less than a decade, maybe two. By then, artists and scammers will have found the best ways to use A.I. - as did the artistic or scammy scribes, printers, wood-carvers, cartoonists, photographers, Photoshoppers...

The genie's out of the bottle, so how should we use its power?

[Fortnight]

AI is still far too much in its infancy to be properly intelligent enough to do most creative things decently, and the one thing that's lacking in order for it to create art is less to do with intelligence and more to do with other aspects of humanity, and that's opinion, preference, taste.

It's all fine and well AI being "I" enough to do smart stuff, but when will it know what it likes? That's a whole other ball game. Your taste is fashioned from personal experience over your lifetime, the things that were around you at various stages of your life, the limitations of what you could do, see and find in your locality, and the other 'intelligences' around you at those times and what they were experiencing. Your taste is formed from what you couldn't experience just as much as what you could. AI going to struggle at that for a long while yet.

Until then, this half-arsed AI we've got now is just going to get in the way. Intelligence without humanity is just not going to cut it with art.

To continue my previous analogy of digital inbreeding, society needs to regard it as being as distasteful as actual inbreeding in order to halt it in its tracks. Once the novelty wears off.

[The Legendary Shark]

I don't think A.I. will ever be truly intelligent because it lacks the biological component that gives rise to the chemically induced emotions that lead to intuition, empathy, creativity, etc. All it will ever have is the cold, emotionless, logical side. 

The way I understand it, when I am writing a sentence, I have an innate feel for what word comes next and many sentences arrive fully formed in my head with the correct words already in the correct order. An A.I. constructs its sentences one word at a time, knowing only which word probably comes next. I presume the art works in the same way, the A.I. "artist" knowing only which pixel probably comes next, so small errors degenerate into extra digits or missing limbs. The A.I. doesn't have the intelligence to step back and examine its own work or notice obvious flaws. 

I'm thinking it might be very good for generating textures and such.

[The Legendary Shark]

In my attempts to find a use for a.i. I decided to try and reproduce an image I've shared here before; first as a crappy sketch in an old challenge and then as a crappy digital painting based on it, like so...

So, with these here past monstrosities in mind I first generated a load of images at perchance.org like this...



I then opened the GIMP and applied what can only be described as a shed-full of faff to end up with this brand new a.i. assisted monstrosity...

[The Legendary Shark]

Tweaked...

[JayzusB.Christ]

Speaking as someone who makes a living out of drawing and painting, AI art is deeply worrying .  Fortunately mass-produced AI systems can't climb ladders and paint murals yet.  They can, terrifyingly, make massive stickers with similar content within minutes.

What you might see as a fun and whimsical way to illustrate throwaway jokes, Sharky, I see as another step towards the complete redundancy of all the skills and techniques I've spend decades trying to learn and hone, not to mention potential unemployment.  That's how technological progress works, I suppose, but I don't have to like it.

[The Legendary Shark]

I understand your concerns, JBC. Because I've spent a significant portion of my working life as a professional driver I view a.i. self-driving vehicles in a similar way. Fortunately, in my case that doesn't matter much as I don't do that job any more (though it's a skill I have in reserve) but if I was just starting out again I'd be worried too.

Furthermore, the a.i. glitch causing figures with too many limbs or distorted faces in art might lead to rather more significant mistakes on a motorway or high street. Driving, just like painting, is at least as much about the human factor as the technology. There is art in everything we do; and a.i. will not be able to simulate that for a long time yet. But it is coming. Indeed, the hundred-years-hence version of you might own a couple of second hand art robots, each one programmed with the unique owner's style, which do jobs for him all over the place while he reserves his own work for customers willing to pay a premium for the human touch. Similarly, that future version of me might own a couple of a.i. vehicles and reserve his own skills for similar higher-paying clients.

For the present, I find this technology to be a great boon to my creativity. Take the image above as an example and compare it to its predecessors. It's still not very good but it's far better than the originals and closer to my initial intent. The a.i. technology helped me to do this in three major ways; firstly as a mood board, secondly for throwing up ideas, and lastly to generate elements.

As I've droned on about before, I have aphantasia. This means that I can't conjure up images in my head. My mind's eye is blind. Now, I obviously don't know how you do it, JBC, but I have gleaned that some artists see the images they want to create in their heads and "copy" those mental images onto paper. In my head, there's nothing to copy - only a concept; in this case, the wreck of a massive war machine rotting away in the noxious atmosphere of Nu-Earth. The first image I see is the first image I draw. The a.i. I use allows me to generate loads of images of futuristic battlefields, for example, allowing me to explore the mood and composition of the piece. It's a time-consuming and enthralling process because few of the generated images are worth saving for reference.

The above process also throws up bits and pieces of interesting ideas - things that would fit into the concept or suggest elements I'd missed or not thought of. It's also good for fleshing out more abstract ideas. Even in the first image I had some vague notion about how the light from the wormhole in the sky would work, scribbling it out with red ink. In the second, I made the light around the wormhole fiercer and experimented with sunbeams. In the third image, I used a.i. to explore different kinds of wormhole and waded through all sorts of generic and mundane sci-fi t.v. clones before hitting on the idea of a rainbow corona. This is when the concept of Nu-Earth lighting crystallised somewhat in my mind; if the wormhole gave off rainbow light, then the surface of Nu-Earth must glisten with rainbows. I haven't been able to properly express that in the above but I think it's close enough to give the gist and I like it, though it's something I could never have envisioned in my own mind.

Finally, I used the a.i. to generate the major elements for my masterpiece; machinery, wormhole, figures, buildings, wreckage, textures, etc., and used them as I would photographs in previous digital work, cutting and pasting and stitching them all together.

Lastly I started painting and smudging over the elements, just as I would with photographs, and drawing it all together with layers and filters and all the Gimpy tricks I've learned over the years. I am fairly pleased with the result.

So, for me a.i. art is like a prosthetic mind's eye - it still can't do everything yours can but for me it opens up a door to worlds of invention - if not for producing pictures of Nu-Earth for my own entertainment then at least as a tool for exploring visual ideas for my writing projects, where the a.i. images are entirely behind the scenes like the rest of the research.

[JayzusB.Christ]

Now, I obviously don't know how you do it, JBC, but I have gleaned that some artists see the images they want to create in their heads and "copy" those mental images onto paper.

Not me, I'm afraid - how comic book artists do that is beyond me.  For a supposed professional artist I'm not particularly creative.  When I'm commissioned, I have a very, very sparse and vague picture of the whole scene in my head, then for the components I turn to Google for similar images to copy.  And even then the scene rarely ends up as I imagine it. Maybe I have a small dose of aphantasia too, who knows.

[The Legendary Shark]

Well, it is a spectrum. At the far end is hyperphantasia, when people can't distinguish real from imagined images - which must be terrifying.

[Funt Solo]

It is terrifying.

[Jim_Campbell]

What you might see as a fun and whimsical way to illustrate throwaway jokes, Sharky, I see as another step towards the complete redundancy of all the skills and techniques I've spend decades trying to learn and hone, not to mention potential unemployment.

This may have been mentioned before.