AI art in compos
category: general [glöplog]
heyo
so AI art is a thing - it's becoming pretty commonplace and it's flooding all over instagram, linkedin, etc.. it's also started making appearances in demoparty compos.
At TRSAC we are not allowing AI entries in the freestyle gfx compo, since we (or in this case mostly it's me, since I'm the graphics compo wotsit for TRASC compos) deem this compo to be about humans showing off their art-making craft.
However, we acknowledge that there is also a craft in creating great prompts for Midjourney/Dall-E, so we have chosen to create a new compo, the ARTificial (theme) graphics .. theme to be announced as we near the party.
But what's your take on this? If you are a party organizer, are you going to do the same? Or are you going to allow AI art no holds barred in freestyle graphics?
one love
so AI art is a thing - it's becoming pretty commonplace and it's flooding all over instagram, linkedin, etc.. it's also started making appearances in demoparty compos.
At TRSAC we are not allowing AI entries in the freestyle gfx compo, since we (or in this case mostly it's me, since I'm the graphics compo wotsit for TRASC compos) deem this compo to be about humans showing off their art-making craft.
However, we acknowledge that there is also a craft in creating great prompts for Midjourney/Dall-E, so we have chosen to create a new compo, the ARTificial (theme) graphics .. theme to be announced as we near the party.
But what's your take on this? If you are a party organizer, are you going to do the same? Or are you going to allow AI art no holds barred in freestyle graphics?
one love
Code:
sword, dragon and boobs,
pixel art, in the style of boris vallejo,
compo winner quality
Quote:
However, we acknowledge that there is also a craft in creating great prompts for Midjourney/Dall-E
Is there though, really?
Quote:
At TRSAC we are not allowing AI entries in the freestyle gfx compo
Wouldn't they already fail the "please supply steps" rule in the first place?
Quote:
But what's your take on this?
(Disclaimer: not an organizer)
My take is that in the history of time there hasn't been a single application of AI that hasn't been used to justify removing people from the workforce, that the current surge in generative graphics has been trained on illustrator's work that they didn't agree to provide (basic copyright violation but hey, it's tech, not like anyone cares about it cos move fast and break the world, right) and which will inevitably risk their own jobs to the point where artists are now too scared to publish their own work cos if it's not being turned into NFTs then it's being fed into a massive dataset that companies can then use to generate imagery for free, and that we're failing at the absolute bare minimum level of moral rejection when it comes to this shit because we think it's entertaining, which - if we as a scene go down this path - on the absolute micro scale will result in even less interest from the soon-to-be-dwindling number of graphics artists because why would they bother competing against THAT abomination?
But yeah, graphics compos, sure.
"move fast and break the world"
"we're failing at the absolute bare minimum level of moral rejection when it comes to this shit"
+
"we're failing at the absolute bare minimum level of moral rejection when it comes to this shit"
+
gargaj: at least there is a difference in what people are able to achieve with AI, right? I would attribute that to some kind of knowhow ("craft", though perhaps a generous use of the word) with regards to how to use the AI.
Quote:
it's also started making appearances in demoparty compos.
has it? any examples? wouldn't the "supply workstages" rule prevent this?
Quote:
My take is that in the history of time there hasn't been a single application of AI that hasn't been used to justify removing people from the workforce, that the current surge in generative graphics has been trained on illustrator's work that they didn't agree to provide (basic copyright violation but hey, it's tech, not like anyone cares about it cos move fast and break the world, right) and which will inevitably risk their own jobs to the point where artists are now too scared to publish their own work cos if it's not being turned into NFTs then it's being fed into a massive dataset that companies can then use to generate imagery for free, and that we're failing at the absolute bare minimum level of moral rejection when it comes to this shit because we think it's entertaining, which - if we as a scene go down this path - on the absolute micro scale will result in even less interest from the soon-to-be-dwindling number of graphics artists because why would they bother competing against THAT abomination?
I'm not disagreeing with that .. at all, actually :) I'm just looking at some practicalities re: compo organizing.
And I say this as someone who has been ragily using midjourney to angrily sculpt Putin out of various food items, chiefly cheese, and turn the American White House into melting blobs. However, most people won't be using it to vent their frustrations by turning world leaders into melting butter and toothpaste.
And technically speaking even that impinges on what legit electronic artists with 3D software could make already.
I kind of wish there was a software way to make images undigestible by these various algorithms.
And technically speaking even that impinges on what legit electronic artists with 3D software could make already.
I kind of wish there was a software way to make images undigestible by these various algorithms.
Quote:
gargaj: at least there is a difference in what people are able to achieve with AI, right? I would attribute that to some kind of knowhow ("craft", though perhaps a generous use of the word) with regards to how to use the AI.
Honestly all the "difference" I've seen is in which service you use, and those services usually just vary in terms of computing capacity allocated (which is why some of them are invite only). The more capacity you throw at it, the better quality it produces - but you're talking about the end users here, so it becomes almost like a "web search competition" where everyone's using Google because it's that much better than anything else, and the only skill involved is varying the search terms.
Quote:
has it? any examples? wouldn't the "supply workstages" rule prevent this?
well off the top of my head, TRSAC, Black Valley, Nova
re: workstages - it depends. With Midjourney, image creation is an iterative process where you start by prompting the AI and then choose options to guide it in the direction you want to go. The steps along this iteration have been submitted as working stages.
Quote:
Honestly all the "difference" I've seen is in which service you use, and those services usually just vary in terms of computing capacity allocated (which is why some of them are invite only). The more capacity you throw at it, the better quality it produces - but you're talking about the end users here, so it becomes almost like a "web search competition" where everyone's using Google because it's that much better than anything else, and the only skill involved is varying the search terms.
Which may not hold a lot of merit, sure. Phrased like that it certainly doesn't :)
Dude, its cheating to use an AI for a graphics compo brcause the entrant did not draw it.
Yes, this is also my position
Quote:
Quote:But what's your take on this?
(Disclaimer: not an organizer)
My take is that in the history of time there hasn't been a single application of AI that hasn't been used to justify removing people from the workforce, that the current surge in generative graphics has been trained on illustrator's work that they didn't agree to provide (basic copyright violation but hey, it's tech, not like anyone cares about it cos move fast and break the world, right) and which will inevitably risk their own jobs to the point where artists are now too scared to publish their own work cos if it's not being turned into NFTs then it's being fed into a massive dataset that companies can then use to generate imagery for free, and that we're failing at the absolute bare minimum level of moral rejection when it comes to this shit because we think it's entertaining, which - if we as a scene go down this path - on the absolute micro scale will result in even less interest from the soon-to-be-dwindling number of graphics artists because why would they bother competing against THAT abomination?
But yeah, graphics compos, sure.
Short answer: This.
Long and probably still incomplete answer: This, and my personal opinion (while freshly recovering from being a first-time graphics compo organizer) I have this strong feeling that it is, more than ever, up to everyone holding some sort of "bouncer" responsibility to not just figure out what our standards are, but what and who we want to see on the big screens a year from now.
What do we want to encourage?
As an artist (and trust me - calling myself an artist feels wrong, but i'll roll with it), i want to also add that growth happens on the road between "having an idea" or "being inspired by a particular art style" and ultimately landing somewhere where personal skill, the idea of who we want to be as creators, and the secret ingredient of a good dose of side-distraction and "fucking around to find out". Copying is about reverse-engineering not just the art style, but the thought behind it. Is it worth it? Oh fuck yes, cause fortunately, us humans are extremely shit at actually reproducing something without bringing in our "shortcomings" into the game - amplifying those and figuring out what's different about our end result makes us develop the thing that ultimately becomes our signature style.
YMMV but bereaving yourself of that process means that ultimately you won't ever get to the point where you start respecting yourself as a creator, so I guess everyone can decide for themselves if they're willing to do the work. I know i do - everyone who's currently in the business of "giving a fuck" does. And those are the people I want to see, encourage and learn from.
Quote:
Dude, its cheating to use an AI for a graphics compo brcause the entrant did not draw it.
That is a gross oversimplification of the problem here. First off, procedural methods have been used in graphics in time immemorial: even Deluxe Paint (notoriously) had flood fill or line drawing where the artist didn't draw every single pixel, but especially in the Photoshop era you get to the point where an image can consist of a collage of parts of photos, 3D rendering, vector art, clipart, stock brushes, etc etc. - hell even Photoshop's "content-aware fill/delete" is (iirc) based on some sort of machine learning.
"The entrant did not draw it" is a horribly reductive way to phrase a manifold issue. Where does the author's visual creativity begin and end when the entire image is composed and produced by an algorithm? If the algorithm was trained on a massive dataset of existing images, are the creators of the original images (who didn't necessarily agree with the creation of derivative works) also co-authors of the work? What's the quantified proportion of involvement between the person giving the prompt, the person developing the AI, and the data it's being trained on? All these questions have very contextual answers, but in this particular case we do actually have a well defined context: a competition where we expect the entrant to show artistic / compositional / technical skills within the constraints of the competition rules. Does this define "drawing"? In some cases yes, but we've been way past that stage since the mid90s.
Quote:
However, we acknowledge that there is also a craft in creating great prompts for Midjourney/Dall-E
maybe, but judging by the large jump in quality in many artists social media accounts since they started using midjourney, i'd say it's not quite as difficult as the craft of making graphics manually.
Ok, how about we define "draw" as a set of decisions (ignoring the doubtless capacious philosophical literature that may define this concept and indeed the concept of decisions, and whether a non-human intelligence is capable of making a decision, see Searle's Chinese Room concept etc).
In its most simplified sense one can argue that to draw an image, an artist:
- Decides how to compose the image (rather than just traces an image search result from the web)
-Decides where on the canvas to make marks and how those marks should look. In the case of the action painters they are still deciding where to point their brush, and the Dadaists also had to make decisions even if those were free association.
-Decides when the image is complete.
(I'm ignoring stamps and precoded paint effects here, because of its essence this is why in a compo you say what tools you use if you don't just photograph a drawing on paper. One can argue that Procreate is cheating but you still need to know a lot to get a good output from it.)
If you code an image, you still have to make decisions about which marks you place where. An unaltered or traced slice of a Mandelbrot in my mind is no acceptable coded graphics compo entry.
Drawing is not just the physical craft of making marks on a canvas, it is the series of decisions you as a human make to determine the result of the drawing process.
In its most simplified sense one can argue that to draw an image, an artist:
- Decides how to compose the image (rather than just traces an image search result from the web)
-Decides where on the canvas to make marks and how those marks should look. In the case of the action painters they are still deciding where to point their brush, and the Dadaists also had to make decisions even if those were free association.
-Decides when the image is complete.
(I'm ignoring stamps and precoded paint effects here, because of its essence this is why in a compo you say what tools you use if you don't just photograph a drawing on paper. One can argue that Procreate is cheating but you still need to know a lot to get a good output from it.)
If you code an image, you still have to make decisions about which marks you place where. An unaltered or traced slice of a Mandelbrot in my mind is no acceptable coded graphics compo entry.
Drawing is not just the physical craft of making marks on a canvas, it is the series of decisions you as a human make to determine the result of the drawing process.
Isnt it the end result that matters? At least thats the argument been made for making demos using tools like Unreal/Unity/Notch. So using any tools to end up with an attractive end result should be ok here too then, no? Or? Anyone can spot a Midjourney picture, as well as a Notch demo, so it would be up to the user to vote for it, or not.
It’s the creative process and the personal learning that matters.
Putting Notch and AI-generated stuff in the same box is ridiculous.
Putting Notch and AI-generated stuff in the same box is ridiculous.
Easy litmus test: whose string of explicit compositional decisions, which can be influenced by the nature of the medium, determine the result. Notch does not make decisions for you.
the day notch can create a demo for you based on a few prompt words.. is the day i become very, very rich.
@farfar: My take is that the line between drawing and (AI) generated image is extremely blurry, and these tools make it so blurry it practically speaking becomes meaningless. How do you even phrase compo-rules that disallow AI techniques with clear lines? Does this mean we can't use content aware fill, or the repair-filters in Photoshop? Will you provide a list of approved software and plug-ins/filters?
I don't think you can do this in a meaningful way...
I don't think you can do this in a meaningful way...
