Interesting to follow this conversation.
I’ll try to write something pretty developed, but one thing that fascinates me with the whole AI debate: art is so tied to human labor as the main value, rather than the idea. I’m so glad this movement is finally challenging things. Burnout rate among artists is off the roof. I’d personally attack the system rather than the kind of tool revolution currently happening.
I’m all for borrowing machine talents to get more time to dive into references and art history. Humanity can only benefit from overall being more curious and agile with language, more than yet another artist spending countless hours doing pretty pictures. The technical level of artists is at its peak, the creativity, hmmm I don’t think so. A very bold and generalized stance, of course.
My day job is as a computational designer, mostly using both text-based and visual scripting platforms in 3D CAD (I mostly use Rhino and Grasshopper, if anyone knows what those are). Lots of coding in C#. I create a lot of my own algorithms, but also am well aware that I am consuming very advanced API’s, which have insanely complex algorithms underpinning them - geometry kernels are truly extraordinary, and way out of my league. So I rely heavily on algorithms written by others to write my own. Often I’ll create my own implementations of algorithms that others have already written or described in detail, say, in a research paper.
The deeper one goes, the more one realizes that the frameworks in which we create are hugely reliant on the work and creativity of other people. Same goes with electronic instruments. When I build a patch from scratch on my Digitone, the synth engine and constraints of how the UI/UX exposes its parameters are fundamental determinants in the outcome.
The recent emergence of public-facing GANs for image generation are diagrammatically identical: there is a UI/UX framework that deploys visual algrorithms based on input. The process of creating prompts that lead those frameworks to produce compelling images is absolutely a form of creativity. You can argue that it’s “simple”, but the person entering the prompts is not doing nothing. The prompts that went into many of the images in this thread are hard to formulate, curate, and process.
It feels closest to collage to me, or perhaps to the work of artists like Dave McKean. But it’s clearly an art form.
With this logic you should be ditching all your music “gear” [instruments 
being mindful of another thread], and buy a bunch of loops and midi packs, put them in a folder that you can access with a RND generator and call it a day.
It doesn’t have to be all or nothing. I use samples sometimes, is that OK?
I think you’re also dismissing a lot of effort and talent amongst those in this thread - and although you’ll see plenty of both here, there’s a big difference between asking an ai to give you a picture of a happy dog, like you’re doing a google image search, and spending an hour crafting a render based on very specific ideas. Some of the works above are incredible and unique and specific to the minds of the people that helped create them. (note that the difference I see there is not the effort imparted, but the influence of the person in the generation of the art)
But honestly I find the debate about whether or not Ai art is ‘real art’ to be kind of pointless and academic, I should really stop rising to it 
The much more interesting questions are around subjects like ownership, perceived value and how it impacts the professional world, this is good material!
I started feeding “shelter from insanity” into wombo dream. Why did I think I’d get positive imagery?
edit - because every time I read the news, this is what I crave…
That’s what I was on about. Claiming ownership of the AI’s output.
I still think it’s sick! Some of the renders are insane. But I could never say “I made this”.
I got MJ to kick this out.
Even the bite stuff is cool. The Moebius style stuff is crazy.
I think we probably feel similar but articulate it differently. It’s interesting that you remove your agency but then immediately bring it back, i.e. ‘i didn’t do it, I made something do it’ - it’s clear that you see at least some form of collaboration - you still made it happen. Unlike a google query where the results always existed, the output of your AI prompt is something you made happen.
I find that the effort I put in (there’s that word again) impacts how much ownership I feel over the works created, also.
A throwaway query is almost like a google search - but I have some collections I’ve created that I feel quite a lot of ownership over - the amount I influence the outcome impacts how much I think of it as mine - which makes sense I guess.
So that is the one point I’ve been avoiding bringing up. That say, in photoshop you need to use it to make something happen. And nothing would happen without your input.
And I KNOW you’ve been chomping at the bit to have that door open so you could say the AI can’t do anything without input. But you’ve decided to put werds in my mouth to open that door for you.
Yes of course it needs some input to generate something. No matter what I say you will twist it, like, you don’t gain any talent or skills just putting in prompts. And you might reply something along the lines of learning how to be more refined in your definition is a talent. Of course paint doesn’t paint anything without someone pushing a brush. But the more you do it, you learn a skill. And that skill is defined by the person doing it. It has their “dna” in it.
And you bring would most likely bring up the refinement of prompts again. Whether you are telling it to do 3d, 2d, watercolour, oil….you yourself are only doing one thing. Submitting a definition. While the person that paints with oil, then watercolour has learned how to do two completely different things. I can paint with watercolour. But I SUUUUUUUCK with oil. It’s hard to learn. But it would take NOTHING but a werd to do with AI. No skill acquired.
Bottom line is, you yourself are not creating a piece of artwork. You are writing a book of descriptions. The AI or whoever wrote it, is generating the art. You are borrowing talent. I wouldn’t call it a collaboration because there is no back and forth dialog. There’s just you pressing a button or adding a werd forcing a result from an image generator, that you can’t FULLY predict what it will be like. You really are rolling the dice. But when I set out to actually draw something, I know what’s coming. And I have have FULL control over that.
You would blow my mind if the AI replicated, with precision, a drawing you drew, by only giving it a prompt. No upload of your own werk.
Ain’t gonna happen.
Art by three guys I werked with. Infinitely cooler because it came from their hands.
I myself, do not posses such talent.
- Justin Sweet
- Kevin Llewellynn
- Android Jones
Many of the algorithms I use when I design surprise me in their output. Yet I have written the code. I don’t “fully” know what is going to come out of them…that’s part of the point. The algorithms I write and/or implement basically grow into different structures, or transform existing geometry into something else, but at least partly out of my control.
The code I wrote to design and specify fabrication files for that project relies heavily on a variety of both programming libraries that other people wrote and generative algorithms that other people described, and the form itself was emergent through the design process. Is that art, or was it nothing more than prompt curation on my part? Is there a difference?
Many of the images presented throughout this thread are heavily processed: they start with a text prompt that is interpreted by a GAN. That prompt is either refined, or re-run, or translated into another design tool for further modification, like Photoshop, perhaps later to be sent into another GAN for further development.
You seem to insist that knowing what is going to come out the other side is a fundamental component of creating viable art. I would suggest you ask any of your colleagues who created the lovely art you posted if they knew exactly what they were going to create before they did, or if much of their output emerged as they were making, took form as it was being made.
I think you are creating an artificial boundary. It’s been done many times historically, as new media or platforms have emerged in the creation of art. Electronic music has long been denigrated by many musicians as not being “real” music: in their argument, if it’s programmed, not played, it isn’t art. Television was similarly dismissed. Much of modern art was dismissed as being purely procedural.
If you think so much of what has been shared here is reductive and that the individuals curating prompts and sequencing image processing steps aren’t creating art but are merely one of the million monkeys at a typewriter now gamed by AI to write Shakespeare, you might want to have a go yourself at creating some of the visually cohesive narrative shorts that @echo_opera posted here, or any of the other pieces that vary on a theme, with just a “simple prompt”.
The ‘feeding ai’ with field recording to create a new sound grabbed my attention. Do you have some links for DIY ai music creation?
Thanks for sharing - I resonate with your reflection on AI art. It will be quite interesting to see the ramification and development of the tools in the coming year, and it’s evolving in leaps and bound. What you describe as a limitation now will most likely change quite rapidly, and the bias will also change radically, but there will always be bias, until there is a deeper learning algorithm. Similar to Deepmind learning to play go from the rules created by human… then learning Go by playing against itself a million time to develop its own way of playing…
I follow this thread with great interest as it boils down fundamentals of VALUE of Art and Labour in general. First i was highly enthusiastic in trying out MJ and shoot out my free cpu cycles. While doing it, it felt not “right”, but that’s my personal and subjective view on this “creative process”.
I have another perspective on using ai for creation and feeding it in the visual (and soon audio) realm or in using ai for touching our sensoric systems:
The club of Rome released a report on the global challenges we are facing, and quite logical intervention concepts, on last Tuesday. While most (5) threats to us ALL should be obvious and I hope all of us here are aware of. One overall problem is, that these challenges can’t be faced, when we are not able to co-work on these globally due to:
“Because the most important challenge of our time is not climate change, loss of biodiversity or pandemics,” the group said. “The most significant problem is our collective inability to distinguish between fact and fiction.”
This, my forum friends, is a MUCH larger issue with AI then the value and perception of Art.
IMO
Ok so this is wonky but also quite fun You provide a sketch with your prompt.
Blue eyed octopus man with sausage legs posing for a square portrait. Very handsome octopus photograph.
A very sad humanoid banana, but it has no reason to be sad because it is very successful and loved by its friends.
I’m exited about where this AI art may lead. Imagine immersive VR environments created from prompts that we can explore - whole universes on par with huge games like No Man’s sky or similar. But more psychedelic, or realistic, or nightmarish, or, or… whatever you prefer. The prompts will be epic, at least for those poets who know what they’re doing. Then they can invite friends to come play in their universe.
Keep feeding that machine I say!
This is the most interesting post I’ve ever read on Elektronauts - thanks for posting.
@dubmo - can you post a link to the original Club of Rome report that your quote comes from - thanks!
I think you moved the needle to a different track, or maybe not.
I think most people will agree with you that the ability to distinguish fact from fiction is an important skill, (and maybe it’s unfortunate that we’re required to point that out).
That reference, though, in the context of your opener:
Star Wars is a work of fiction.
Peter and the Wolf, narrated by David Bowie (recommend everyone listen at least 1x), is a work of fiction. HR Giger’s fantastical fictional drawings made him famous.
I digress here only to establish a fact is something that is true. There is no xenomorph, and there is no Death Star. They are fictional.
What I’m taking away from your post is a suggested correlation that Fact is to “real art” created according to a set of rules that no one agreed to ever, as Fiction is to anything purported to be art that doesn’t fit that set of rules.
If I see a picture of a banana with a face on it, and it freezes time for 1 second of my life and makes me smile, am I somehow a party to the perpetuation of:
Creative professionals cannot will the public to reject AI artwork as “fiction” any more than vineyards in the Champagne region can throw shade at sparkling wine from California. Ultimately, it’s up to the audience to decide what they desire and what they will accept.
But don’t take it from me.
Check in with the candlemaker who insisted that lightbulbs made fictional light.
How’s that guy doing?
The quote I cited from an article on Welt here --> Club of Rome: Es ist noch nicht zu spät, die Menschheit zu retten - WELT
The original is quoted from the actual Ckub of Rome report Earth 4 All which is yet to be delivered and ordered.
–>
I think you are right about me, sidetracking the issue. In addition, my post wasn’t meant to push or pull on the discussion about “real art” or fiction in a creative, story telling sense. This is minor to my point (but truly fascinating on many levels)
I was turn onto the technical possibilities of blurring and lifting the line between the sensoric reception of facts and fiction in a informative context, not art or writings. The ai is a massive tool for disinformation on a broad, society level which will be / is used for counterproductive and destructive actions. I hear some saying, “yes, propaganda was always part of the game” but we talk here about it’s hyper connected, “max upscaled” version. This leads to unforseen development of controlling the perception of reality and therefore, behaviour of people on a global level.
Maybe this thread about the Art corner of Ai is not the right place for this discussion. In my book, it’s secondary while being also a essential branch of the tree. (someone thought about education in the context of ai and truth/reality…man, this topic is huuge) as I saw this article on Welt, and the quote of the blurred reality as threat to handling the massive global issues, my thoughts about this got underlined and I wanted to share them here.