AI wins state fair art contest, annoys humans
arstechnica.com
arstechnica.com
> Out: Art that makes the public mad by deconstructing the idea of art
> In: Art that makes artists mad by deconstructing the idea of artist
There's space for various forms of art competition. You could have pure competitions with stricter constraints on what is or isn't allowed.
I love that humanity has developed so many new highly advanced tools which allow greater artistic expression by a broader range of people. The creation of the synthesizer didn't kill the guitar, it just opened the door to a whole new world of musical expression.
If I wanted to capitalize on the AI craze that's currently happening I'd host an AI-generated art competition / exibition. I'm surprised nobody from the companies that are developing these AI tools has tried doing that. It would be really interesting to see what kinds of works people are coming up with after investing dozens or hundreds of hours into these AI tools.
I suspect a lot of AI artwork will end up looking kinda same-y after a while, but a few people will figure out how to hack around with prompt engineering to get really interesting and unique results.
[0] https://twitter.com/C_Harwick/status/1564948225781555202
compare music as played in a cultural setting, say 100 years ago. and the avalanche of dance tracks in the last thirty years (ravers, you are not invited to do this test). Honestly tell me that creatively nothing is lost in musical arts.
Compared to the 1800s, art has been pushed towards things that are "simpler" in all areas and it's not because of digitalization. Rothko didn't happen because of a bucket fill tool.
This can't be true for still images. Not only the quantity of art being produced by digital means but the variation and quality has exploded since the democratization of digital painting. Certain styles are in vogue but there are countless examples of wild experimentation. Technical ability also seems greatly enhanced by these tools. Add to that all of the forms that weren't even possible 30 years ago and I strongly doubt we are in any way losing variety or quality in art. Maybe the sheer quantity is the problem, it's harder to find the really great art amongst the ever expanding ocean of generic art.
The digital music you're complaining is just one portion of what is being made and it has mass appeal; you can also complain that sitcoms are so much worse Citizen Kane or McDonald's is so much worse than fine dining used to be. There's still nuanced music and film and food being made, and the complaint that popular stuff is "crude" is basically just a normal refrain in all areas of art, and I don't think digitalization is the primary culprit.
I say this with no snark intended: you are demonstrably in a filter bubble and need to expand your horizons before you can accurately make comments about what "most" art is
> Deviantart is effectively a filter to specific subset of digital art
may be true since it is a community of artists that tend to have their own sub-genres that they move in but I believe the other poster was just giving it as the most popular example. Deviantart is really just the tip of the digital art iceberg, you can search and find a multitude of other sites hosting art communities as well as the sites of individual creators, YouTube and Twitch streams, Discord servers, etc.
The very first OP in this thread is arguing that music was ruined by the availability of digital music compared to music 100 years ago. I think this is just the standard "kids these days" griping. I was trying to say that digital art isn't special, in that most of it is also crap but it's not hard to find good stuff. Music is the same as that and has not been ruined.
Music is more creative and varied than ever. You just don't like it as much.
https://blogs.scientificamerican.com/observations/is-pop-mus...
:)
I have a music TV service and I have to say that today we have very deep and creative music and bland and shallow too.
Technology turned up artists expressiveness incredibly, but a good artist can create wonders with extremely limited tools and a bad one can't create anything good spending god's money on expensive gear. We have good artists today too.
There is a definitely "look" to MidJourney pieces, that once you've seen a bunch, they're somewhat easy to pick out.
That's a bad argument. First of all, it probably did replace guitar parts. Computers with plugins certainly did. Second, the argument was rather that it would replace musicians. And it did that too. Modern samplers replace entire orchestras. Not that it's a bad thing per se (after all, who can afford even a small string section?), but it did happen. Third, this doesn't take the same role in drawing a synthesizer takes in music production.
> I love that humanity has developed so many new highly advanced tools which allow greater artistic expression by a broader range of people.
In der Beschränkung zeigt sich erst der Meister (It is in working within limits that the master reveals himself).
The invention of the synth and samplers has definitely replaced the jobs of many instrumentalists. Despite that, people still learn how to play the guitar, the violin, the trumpet, etc.
However, it also decoupled the roles of musician and instrumentalist so that a wider range of people can express themselves musically without dedicating a majority of their life to learning to play one or more instruments.
AI art tools are going to do the same thing - the roles of 'artist' and 'painter/draw-er/whatever' were previously tightly coupled but you no longer have to have the raw technical skills.
As it was with music, this is probably a net benefit to society since it allows so many more people to express themselves artistically. It will cost some people jobs, though.
It will also deter some people from learning to draw/paint but there will still be plenty of people who do it anyways.
1. Does capitalism somehow obstruct people expressing themselves musically? Does nationalized ownership of production means or being ruled by clergy stimulate it somehow?
2. Is there a world-view that enables (advanced) musical expression without studying?
3. Is it artistic expression if you bang out a few random words on a keyboard and something is generated in return?
4. What makes you think that this (the topic still being AI generated art) has already happened in music? AFAIK, those tools are pretty limited, and barely accessible. If it did, what makes you think that it benefitted society?
5. How does many more people expressing themselves artistically benefit society?
5b. 5, if it doesn't require any effort?
I do not own an organ, but I can get a pretty good VST alternative.
People have been producing electronic music without a lot of formal musical training for decades now.
So an image like this — which by the way, I will admit I find on some level beautiful and interesting to look at — which doesn't have any intention behind it, is thought to probably be be some kind of deep art.
> Still, as long as new artistic tools emerge, the "Is it art?" debate will likely continue as long as there are people to have the discussion.
For me the question is moot.
The question matters deeply to those trying to make art tradeable. Their business is about maintaining scarcity (a.k.a. non-fungibility), so they'll spend up to some portion of their profit margins on gatekeeping and defending terminology.
https://www.theartnewspaper.com/2014/11/01/did-marcel-ducham...
The people up in arms are the critics, the gallery owners, the rich kids who collect items made by people good at making interesting stuff. Because if art isn’t special, if it doesn’t have some deep rooted special meaning that transcends reality. They are not better than the sports commentators and shopkeepers they so desperately try to elevate themselves above.
That said, I think it's uninspired and derivative and does nothing for me. It does seems "skilled" in the sense that there's a decent use of anatomy, lighting, and composition. So on first glance it's a pretty image which might be all someone wants and to each their own.
I would love to see the high-res image though to see if it holds up in the actual details. AI tools seem to have a hard time with actually making sense in the details and has a really hard grasp on things like hands (which are difficult for humans).
All of the arms in the scene are melted stubs. The subject in the middle has what looks like an insect's leg jutting out of her waist.
> I would love to see the high-res image though to see if it holds up in the actual details.
It clearly doesn't. The judges probably thought that there was something cool about the visual weirdness of the piece—which indeed there is. However, it's much less interesting when you know that the melting arms aren't the artist intentionally playing with form—it's just a machine not knowing what arms are.
I think you're right but it's so hard to see in the Ars pic.
A locally famous guy was showing and selling modern art painting by his upcoming protege, it got some traction and interest from buyers and critics.
Then he revealed it was a chimp throwing paint at canvas.
(Especially modern) art scene is the emperors-new-clothes situation + speculation + tax avoidance blended together.
There is something to the "the emperor has no clothes" situation, but also, there is more than that to art.
Except that it does.
> "I have been exploring a special prompt that I will be publishing at a later date, I have created 100s of images using it, and after many weeks of fine tuning and curating my gens, I chose my top 3
A human made the '100s of images', and finessed it down to get what they're looking for. Then, this human picked their favorite. Presumably the human also named this. There's plenty of human intention here.
The headline is clickbaity; a human won an art contest with a piece of art that was the output of a computer program.
Yes. My CEO has excellent technical (and business) knowledge. He will often give me direction on my projects because he knows both the business requirements but also the technical implementations that might fulfill those requirements.
His input is very valuable, but to say he is a programmer or that he is creating these programs is wrong. He hasn't written a line of code in many years.
Being an Art director is a management position. Having domain knowledge is surely useful, but you're managing, not creating.
Are you sure you know what an art director does? They might or might not be the line manager of the artists but that is not their main responsibility. To quote wikipedia: “It is the charge of a sole art director to supervise and unify the vision of an artistic production. In particular, they are in charge of its overall visual appearance and how it communicates visually, stimulates moods, contrasts features, and psychologically appeals to a target audience. The art director makes decisions about visual elements, what artistic style(s) to use, and when to use motion.”
I would call that creation.
Sure they don’t just grab a brush and start painting. (Except when they do, because that is the best way to communicate what they want from others.)
Would you also say that movie directors are not creating just managing?
I don't see how you can call it creation when they don't create anything. They make decisions about what to create, they provide input on how to create, but they do 0 creation themselves.
If I ask my partner to cook dinner, does that mean I made dinner?
> Would you also say that movie directors are not creating just managing?
Yes. It's in the name - they are directing. They are managing the creation of their team. They themselves just sit in a chair and bark orders - not creating anything.
The page was blank before they showed up.
>If I ask my partner to cook dinner, does that mean I made dinner?
Did you just ask "Cook dinner" and wait or did you stand behind them and tell them exactly the measurements and process and throw it out every time it didn't taste right?
https://www.britannica.com/art/auteur-theory
For art predating films, see also: Richard Wagner and his concept of Gesamtkunstwerk
No, but I wouldn't give them all the credit for the acting, set design etc. Which is what the guy in this article is claiming - when his picture is patently 99.9% the skill of the creators of midjourney, plus the millions of EXISTING artworks that were fed into it, to teach it what art is.
Being an artist is a fusion of the creative impulse and the detailed creation of that work. typing a few words does not constitute detailed creation.
At no point did he claim to be a genius, or a great artist.
He was perfectly upfront about his tools. He even posted about his full process before results were announced.
His only mission is to get a new 'AI Assisted' category in competitions.
As has been pointed out, often, many artists came up with concepts and then directed assistants to create based on a brief. Courts have weighed in and found that the original artist owns the work. All that has changed here is that the 'assistants' have been replaced by AI.
Your argument seems to be that only famous and established artists, "masters", ought to be allowed to do this - am I reading that right?
(FWIW I have no dog in this fight, just think these are some interesting questions around the very fuzzy border of art/not-art. I don't expect to resolve the issue here! I'm genuinely just curious about your opinion.)
I wasn't familiar with Sol Lewitt, so looked him up - apologies if I miss the particular aspect of his work that caused you to pick him. He seems like an interesting artist - I like his wall paintings.
If I'm right, he created diagrams / instructions for the production of the wall paintings, and didn't paint them himself, so that's a deviation from what I said about an artist needing to create their own work. However I think on balance, the overwhelming majority of the creative work involved was still his.
One thought I had, was to consider what the difference would be if the artist was removed from the project, and it was completed without their input. Without Sol Lewitt's diagram, nothing comparable to his work would be made "by chance".
Contrast that to a DALL-E user. DALL-E could happily spit out great looking images all day, with prompts generated from random noise. It ONLY makes (reasonably) good looking images. If 9999 out of 10000 images were random noise, and only certain arcane prompts would create a masterpiece, then I could accept there was skill in hunting for the prompts. But as it stands, it's more akin to being a buyer of art in a vast gallery, in my view.
Hope this makes sense!
"Art" is a very low thresh-hold to meet. "Interesting" or "high-effort" is a higher threshhold to meet. Art doesn't need to be good to be art.
Yes. This has happened frequently across the history of art.
Would you not call verrocchio an artist?
https://en.wikipedia.org/wiki/Exit_Through_the_Gift_Shop
Is he an artist, or an entrepreneur taking advantage of the art scene? It's hard to say, the result is quite fantastic in my opinion - it's like postmodern art that unintentionally mocks itself by being based on "stolen" ideas.
So either it's "real art" by expanding our conception of what "art" and "artist" mean; or it's "pretend art", only an imitation of art. I mean, modern art pretty much deconstructed itself, I'm not sure if there's even any difference between authentic ("natural" and "human") art, and inauthentic ("artificial" and "machine generated") art.
Théåtre D'opéra Spatial is a beautiful work and really draws in me in.
If that hypothesis becomes true, I think it will have severe implications for the value of human intention in art. Would we prefer that endless feedback loop over the painstaking labor of honing a skill for oneself? I don't think it's clear yet, because the technology is still too limited to test such a hypothesis. But would that become the norm if it were possible?
Of course, I don't know if such an subjective concept is encodable in an algorithm. But I'm not beyond believing that such an advancement would be highly desirable if the technology progresses far enough, to the point that something "good enough" for most people and aesthetic styles will emerge. People have already been taken by surprise from the developments in generative art so far.
This means, the light patterns that we call image mean completely different things depending on who created them. The art critics or curators don't simply look at the image and understand something from it, they know the artist they know the trends they know the word on the street and they try to understand how the artwork fits in all this. The same straight line image made by Picasso, made by Dali, made by a toddler and made by AI would have completely different meaning and value.
Most tech nerds have no decoders to decode artistic meaning and as a result they will look down on not only critics but also artists. Sometimes meaning can be conveyed by analogies but for that the artistic counterpart needs to special effort.
I once was in a modern art museum featuring an enormous room. In one corner were 3 beach chairs.
That was the art. The curator encouraged me to think deeply about the piece, its meaning, its intent, the artist behind it.
Yes, I'd call that snobbery.
To be fair to the AI, I believe it was trained on both real life and other art of high visual quality - I wouldn't be surprised if it has actually encapsulated deeper meaning without the developers intention to do so.
> So an image like this — which by the way, I will admit I find on some level beautiful and interesting to look at — which doesn't have any intention behind it, is thought to probably be be some kind of deep art.
Beauty is in the eye of the beholder, and I believe meaning is also the same. Many poems for example have been misunderstood from the poets original meaning, but that doesn't stop it being meaningful. Music is much the same.
Just because the AI didn't have intention when creating the art, doesn't mean that it didn't capture it. Nature didn't have intention of deep meaning over millions/billions of years, yet we can take a photo that has deeper meaning than the subject matter itself.
With the photo you may argue "the artistry is from the composition process", but one can argue that the person generating the AI art also went through a selection process, and chose three pieces that they found compelling.
I don't think AI art being chosen proves anything about critics. We have, also, not seen the competition either, and cannot come to _any_ conclusions about the result.
Midjourney/Stable Diffusion/DALLE are doing to Photoshop what it did to traditional drawing methods. But there's still a human in the loop.
Let a monkey write for a million years and choose the best writing. I think that currently AI is better than that, but not so far as to be a misleading comparation.
A lot of the "humans will still be needed" arguments implicitly rely on the assumption that the trend of AI progress stops here.
The original DALL-E was released just a year ago. In one year, AI image generation went from generating images that looked slightly normal if you squinted really hard, to the beautiful, detailed and creative results we see today.
The same speed of progress can be observed in almost all other domains, from code generation to question answering and music synthesis.
For many of these problems, it almost feels like the next step up in quality is just two hours of planning and a couple months of training away. Like they're already solved and just waiting for someone to implement the solution.
So yeah, humans are still needed right now. But I wouldn't base my hopes for the future on that, because the world of AI capabilities will probably look very different in just one year.
Conversely, even if something IS made deliberately and by a human, I'm not forced to recognize it as intentional. I could likewise argue that the guy making it is just a big blob of molecules following (mostly) deterministic rules, and deny them personhood. It would be a pretty shitty thing to do, but it wouldn't be factually wrong. Or right. As I said, it's a teleological choice to ascribe something the ability to have intentions.
But it would not make much sense to ascribe intentionality to both a computer algorithm and the person using it, seeing as the former is a product of the latter.
However I think he was trying to get this story.
I am spending far too long on OpenAI and I see it as next level creative tool. It is not a gimmick like that deep mind Python app that Google had a few years ago (that turned everything into an uncanny dogs head).
What I like is the confidence this new technology has. I have seen lots of things go digital over the years and always found an artist backlash. Digital photography, even Photoshop itself. You can create the next Mona Lisa with nothing more than lasers and space dust and an artist will snub you and go back to making their little clay pots.
With this new OpenAI the results are compelling and there is no denying it.
You also don't need to know any language apart from English. It is not like you need to do something like know how matrices multiply or do trigonometry (yep, that basic stuff we learn at school and never use in real life).
With any tech that is 'wow' you have to make the tech not part of the product. Imagine you have a new image viewer that can give you infinite zoom. You use it to show a gallery of blueprints from a museum. Although the new zoom tool enables you to show off these blueprints, you need to just make it utterly understated and not advertised as 'wow infinite zoom!'. You have to keep focused on the content, the blueprints and making the tech transparent to the objective.
With OpenAi it is the same deal, we have the gimmick of fooling the experts. But then when used day to day then it is just another tool to tell the story and to craft the art.
Also funny how everyone reduces art to painting or even worse digital illustration. Talk about 'AI killing art' once the computers create something superior (not a matter of 'subjectivity') to Amore e Psiche.
Thankfully doesn't apply to people who are familiar with the works of Da Vinci, Michelangelo, Picasso, Klimt, Dali.
At the same time I'm painting the macro picture where most artistic labour is of a commercial nature. Where output is leading as long as it is good enough to serve the purpose: sales, entertainment, decoration, whichever.
Apparently, this output, in whichever way it was constructed, was favored. Maybe it's just pretty, maybe it emotes, maybe it communicates something very well. I don't know, but I assume the jury was competent in being an art judge.
My point is that this redirects a lot of talking points by artists to the trash bin.
Art only has meaning when created by a human. NOPE. An artists' "lived experience" is the main driver of art. NOPE. Art is meaning and only humans can communicate meaning. NOPE.
All these arguments are largely sentimental as actual output seems to be leading.
I can already foresee elitist art critics pondering for days over the meaning of a piece, whilst sipping their wine. Whilst "the piece" is a guy in a basement typing "pretty fantasy landscape, epic lighting".
At a state fair?
According to their website, this is the bio of the Exhibition Judge 2022 (they only show one judge but maybe there are more?): "Dagny McKinley is the Executive Director of the Colorado New Play Festival. Prior to that, she served for four years as Director of Development and Communications for Steamboat Creates. She is an author and photographer with five books to her credit. Her photography has been published across the United States. She earned an MFA in Creative Writing from Naropa University. Dagny currently serves on the Mineral Springs of Steamboat Springs Protection Committee with the Tread of Pioneers Museum, is on the Colorado Business Committee for the Arts advocacy committee, and is determined to see a performing art center in Steamboat Springs come to fruition. Dagny believes that anything can happen when you believe the impossible is possible."
https://coloradostatefair.com/competitions/general-entry-fin...
Gatekeepers, curators and prizes are marketing tools, not necessarily relevant to the value of art. They are necessary for promoting art creation, that’s where their value is. They are not relevant, IMO, for attributing value to art.
I just don't have a good description of what it is.
If that quality generates value for a person depends on what weight that dimension that was used as quality metric matters or affects the viewer.
If you have a clear perspective of what quality dimension a judge is analyzing and it matches your own attribution of value, then I say, in this case, the judgment is a good proxy of value for you. But this is in individual cases, not as a statement of the intrinsic value of said piece of art in general.
I am using this writing to think, so it is not a formed thesis on art value and quality. But that's my perception of the world.
But I'd also append what you said with: human individuals probably share a lot the same dimensions of interest (and probably close to the same weights). Plus, those dimensions/weights would probably become more refined as you gain experience evaluating+generating art.
If there is such a thing as "quality", the key question would be whether professional art critics and experienced artists are any better at finding it.
Not saying humans are the only thing that can produce art, just that -- in this case at least -- human artists carved out the space the model could generate within (which is a massive accomplishment).
I have a hard time imagining the same kind of mythology building around most AI work, at the moment, largely due to the low marginal cost of production.
In the case of the state fair, the judges didn't know that AI had participated in the construction of the art. I would argue that a fair evaluation would require being honest about how it was made. It may have resonated with the judges because they were impressed that a human being used a digital tablet in Photoshop to produce these pixels, the implicit assumption they made when they evaluated it.
Even if there is human ingenuity involved in creating visual hints and writing prompts, a judge will need to understand that process in order to properly appreciate that. We can advocate for philosophies of art that consider only the final product, but this doesn't mesh with how people empirically relate to art.
There are scenarios where input may matter, such as in an art contest. Where the craft part is tested, not just output.
In the vast majority of cases though where the public meets art (websites, movies, games, magazines), I don't think anybody cares about how it was made. It's output-driven. It's not art for art's sake, it's art as utility.
Imagine your favorite writer. They have an incredible imagination and writing skills. But they can't draw or paint for shit. Now this writer uses their imagination to generate incredible illustrations to enrich the book. Illustrations that are relevant and beautiful.
Does it matter that AI helped with it, if it got the job done? I don't think so.
Isn't it photography that prompted painters to explore other forms of painting like impressionism (eventually with great success)? I wonder what the current wave of technical innovations might trigger in terms of artistic expression...
I think this settles the matter. It's not as if the camera or the AI registered themselves to the art contest and submitted their work. A tool is a tool. Of course we can limit which tools are allowed, as there are car races, horse races and bicycle races. Maybe there are no Photoshop art contests.
https://www.reddit.com/r/StableDiffusion/comments/x209sb/pre...
https://www.reddit.com/r/StableDiffusion/comments/wzfohj/wip...
(Photoshop I see as a separate issue; people take it better because it still resembles traditional methods to such a large degree.)
If you've played with any of these tools for 5-10 minutes, you'll know what to look for to quickly identify purely generated images.
Have a look at some of these people https://thispersondoesnotexist.com
After flicking through 10 or so of these, you'll start to become more sensitive to the subtle inconsistencies in symmetry, colors, and the unnatural way high contrast transitions take place, etc, and it's the same with anything that comes out of any of these art generating tools. With time, those particular aberrations can be smoothed over, but I think it's kind of like how when we saw Jurassic Park in 1995, lots of people were blown away about the realism of the dinosaurs, but when you watch it now, they just seem like plastic toy figures. Over time, we've become more sensitive to recognizing computer generated animations. Subtle issues with lighting, hair, shadows, etc. When new technology comes out, it takes some time to recognize how to spot these quirks, but over time it becomes pretty obvious.
I see some amazing artists using midjourney etc, and then clipping together fragments and painting over them using classic digital art tooling, with some astounding results. As with any artistic tooling technology, the real winners are going to be the artists that learn to embrace the tech.
I've spent tons of time with a wide variety of tools/models, and I disagree. It's easy to spot when anomalies show up, like yes, this weird swirl on someone's ear is clearly a DL artifact, sure; but the images that people will actually take and use elsewhere will be artifact-free (unless that's the purpose of their usage).
Take a look here: https://lexica.art/ I would bet anything you could take a handful of current generation stable diffusion-generated images (handpicked for the highest quality, no visible artifacts), mix them with human-created images in shared genres, and in a controlled experiment people would do no better than random in identifying their sources.
Truer words have never before been spoken. However, I think it applies more to the tech industry than the Art world.
For a weightlifting competition, those constraints are things like "don't use steroids" and "don't use an invisible string to assist you in lifting the weights".
For an art competition, the constraints might be things like "don't use AI to generate the image" or "don't hire another artist to do the work for you".
In all of these cases, breaking the constraints is still work. Using steroids is dangerous and may be illegal. Hiring another artist is time consuming and expensive. Getting good results from an AI model takes many iterations and a lot of tweaking.
If you just care about the result then it doesn't matter how the heavy lumps of metal were lifted, or who created the aesthetically pleasing artwork. But if you care about the competition, then these constraints are important.
This is a good point. But...
1) Was that actually explicitly stated as a constraint in this specific art competition? If not, then no (explicit) rules were broken, I guess.
2) Unlike rules such as "don't use steroids" (which can be enforced through testing), it might be more complicated to effectively enforce rules such as "don't use AI to generate the image": even if you assume some perfect "watermarking" process (that would mark an image as "AI-generated" vs. "human-generated"), nothing prevents a person from indirectly using AI in the generation of digital art (e.g., prompt -> generate image using AI -> paint over image using pbrush.exe, deleting all watermarks in the process).
My point: though you make a fair point, rules that cannot be effectively enforced can be seen as somewhat unfair (since they only affect honest people).
I expect the competition contains some rule along the lines of "entries must be the entrants own work" or similar. So yes, most likely.
There's unlikely to be a rule specifically against AI in the same way there's unlikely to be a rule specifically against getting help from space aliens. Both probably seemed as likely to be an issue up until a few years ago (especially to people who don't follow tech developments).
> 2. Unlike rules such as "don't use steroids" (which can be enforced through testing), it might be more complicated to effectively enforce rules such as "don't use AI to generate the image
My understanding is that testing for all banned substances is actually extremely hard.
In any case, the point is moot - this artist did use AI and is not trying to claim otherwise.
If you can get away with breaking any rule, whether it's steroids, AI, or invisible strings, then that's a different discussion.
What exactly counts as "own work" in the age of digital tools and AI? This was explicitly the digital works category, and even Photoshop makes use of AI.
This seems too vague/ill-defined. In what way is the entry "not his own work"? Using tool X or Y in the generation of digital artwork doesn't make the work belong to the tool.
By this logic, using stuff like Photoshop's context-aware fill (which, in some sense, generates/hallucinates novel image data without human intervention) would also make the resulting work "AI-tainted" and "not human-generated", right?
> My understanding is that testing for all banned substances is actually extremely hard.
...but still quite doable (so there's always the risk of getting caught).
How exactly do you propose that digital art gets tested for "use of AI in its production"? Sounds like a situation that is ripe for selective enforcement (i.e., the only people that are going to get caught are the people that are honest/dumb enough to get caught).
> In any case, the point is moot - this artist did use AI and is not trying to claim otherwise.
It may be moot in this particular case, but it's not moot in the general sense, since most people are not necessarily honest (unlike this person).
Either way, unless the competition had an explicit rule against the use of AI tools (or whatever you want to call them), I still don't see how this person did anything wrong: as far as I can tell, the person submitted their own work (note: there is no legal mechanism under which a machine can claim authorship over something).
It's up to the judges to decide, no doubt. Every possible rule (including laws in a legal system) will at some point encounter edge cases. That's why we have judges.
In this case, I expect most judges will feel that having an AI generate art is not the same as having a human draw art using a drawing program, which seems reasonable.
In the context of an art competition, I can see how it would be useful to restrict the use of such tools (as you state), but then it should be done explicitly, rather than through generic/vague "you should make the work yourself" clauses (and, again, it might not be possible to effectively enforce such rules).
This is a good point. The constraints seem important, yet at the same time no one seems to be following them.
Whether it’s elite CrossFit, bodybuilding, powerlifting, or Olympic weightlifting, people ARE taking gear to stay competitive. I could see digital art becoming a similar arms race.
You could probably retroactively disqualify this person from the contest, but it has a lot of implications for the art world outside of state fair competitions.
What we are seeing is not the death of art, but a flaw in our social system, with the livelihood and dignity of individuals being directly tied to their functional importance to society, like cogs in a machine. That is what I think we need to attack.
For all things art, I hope A.I. will help us divorce from the insidious form of commercial art and leave us time to focus on actual art that usher positive social change. An interesting take on that is in Marcuse's One Dimensional Man.
So we are just fine with 'empty' aesthetics? (that is disruptive) So the new definition of art would be: 'a human/ai product of any kind, be it remix or whatever you can sell as art'.
However, it looks like an oil painting, but I think it's a print stretched on canvas like you see in coffee shops. So it's a digital creation.
Whilst not any less of a way to create art. I like classical mediums, including photography (not digital). I think the physical human work of making the art is an important part of what makes something art. I don't discount digital art at all. I just like the older mediums. Anyway, that's just my opinion and I think making nice pictures is a good enough reason to do anything. The creation is the point.
Anyway, good for the owner of the AI who created it?
But again, It's what I like. Some will agree, some will not. I'm happy with what I like and see stacks of digital art I just find myself more drawn to non digital stuff. It perhaps is a side effect of working with tech all day.
I haven't been all that impressed with much AI art until I started seeing these Midjourney images, which I do find compelling for some reason. Which makes me wonder if an AI can have a style for lack of a better word. And then - could one AI be used to imitate the style of another? Like I said, there is much that could be thought about and written about.
From an "art" perspective DALL-E, Imagen, Midjourney bring little if anything new to the table...Not to say they aren't incredible, but "making art with words" has been well-explored for a century: Duchamp's "Unhappy Ready-made" in 1919 is one example, Moholy-Nagy's "Telephone Picture" https://www.moma.org/collection/works/147626 is another. Indeed are these AIs any different from conceptual artists outsourcing their work to other humans? I don't think so. What I _am_ seeing from various conversations, though, is that they are game-changers when it comes to inspiring people to create, and to read and think about art history and philosophy. And that has to be a good thing.
"All items entered for the competition must be entered in the name of the person who created the entry."
If Allen had hired 100 humans to create images based on his prompt, and then submitted his favorite of those under his name I don't think anyone would consider that legitimate. IMO having Midjourney generate 100 images is no different: Allen did not "create" this art, it was created by Midjourney (and by proxy, by all of the artists it trained on).
Olga Robak, the director of communications with the Colorado Department of Agriculture, confirmed that Allen mentioned Midjourney in his submission statement.
According to fair rules, anybody can file a grievance against submitted items — but they will need to post a $300 bond, cite specific rules that have been broken and present a grievance letter in person.
The bond can be returned if the grievance leads to a violation of the rules but Robak said a preliminary review showed Allen had not broken any."
— https://eu.chieftain.com/story/news/2022/08/31/ai-painting-w...
He said (in the Midjourney Discord) that when submitting he "stated that he made them via Midjourney", but it sounds like he didn't make clear that this was an AI tool.
Sincarnate — 26/08/2022 Yes the description clearly stated I created them via Midjourney starshadowx2 — 26/08/2022 Was it explained what Midjourney was though Sincarnate — 26/08/2022 This is why I did it. Should I have explained what Midjourney was? If so, why?
He also mentions that he originally mentioned Midjourney in the caption, but didn't correct them when they edited that part out.
I think these sorts of details do matter.
It would be like having a cake making competition — if you don't want people to use ready-made royal icing or marzipan, you specify that all elements need to be made from scratch.
You name, and explain, what tools you used when submitting an art piece. The general public get to see the "via MidJourney" tagline, and the judges get to have a bit further read of the description of how it was made, during the submission. Nothing here was hidden.
And, if we ignore that this gallery was for digital art, you submitted pencil, then you would say "in pencil", and you would describe the place, and who was present, when you were making your piece.
This is kind of what po-mo theory was about, where people can't tell the difference between the real and represented, what a copies of a copy are, and the relativity of individual experience. AI art is a kind of perfection or demonstration of some of that criticism. What AI art does is use recombinations of existing inputs to produce new symbols that are internally consistent with themselves, and they are only meaningful to us because a) we have accepted the input symbols as premises, and b) we have trained it on our beliefs about consistency that reinforce our beliefs about those symbols. The product of these models not so much "real" as they are just iterations over a finite symbol space. They are artifacts of the logic of an idea.
However, it may be more powerful than we expected, as it's also a perfect meta-criticism for evaluating critical theories, most of which originated (ironically) in literary and art criticism. This dynamic is why you can use ML to produce papers in "theory" that are indistinguishable from acadeimc writing, because these, too, are not honest signals of elevated human thinking or skill, but rather, just self-referential, interally consistent artifacts of ideology whose only bearing on human experience is what we assign to it. Like AI art, they are not an indicator for reality. We can use them to model reality, but confusing the model with the real is full of some pretty subtle tricks.
Anyway, it yields a pretty funny tautology that shows the limits of both art and theory, where arguably, to coin a deepity, if you can simulate it - it's not real.
I would love to print it and hang it in my office…
I don't know whether there's more context for his quip that makes its meaning more explicit, but it looks to me as if it means: the painter, unlike the camera, is not limited to neutral depiction of actually visible things; a painting can be fantastical (showing some imagined heaven or hell) or can convey emotion as well as literal photons (enabling the viewer to share the artist's own internal heaven or hell).
lioeter's comment focuses on the first of those unphotographic possibilities, and as far as that goes it's clearly correct: AI systems like DALL-E, Midjourney and Stable Diffusion are able to make depictions of things that don't actually exist, so that limitation of the camera has been transcended.
But the same is happening to the second too. Mundanely, if you ask DALL-E for "a man standing on a bridge, expressing ..." you get quite different (and, I think, appropriate) results depending on whether you finish it with "horror, dismay, and misery" or "contentment, excitement, and joy". But that's also true if instead of "a man standing on a bridge" you put "abstract impressionist painting of a city". These systems do in fact know something about mood and atmosphere and emotion, in the sense that if you ask them for joy or misery they do produce images that communicate some degree of joy or misery.
(At least sometimes. I tried asking DALL-E for "abstract impressionist painting of a man on a bridge by a fjord at sunset, expressing ..." and I can't say it made the "joy" ones any more joyful or less miserable than the "misery" ones. Same with "post-impressionist oil painting ...". I tried the same with Stable Diffusion and it also didn't really manage. Score one for Edvard Munch.)
They don't seem anywhere close to being able to do that as well as human artists can, though maybe someone more skilled with prompt engineering than I am could get much better results. But they're heading in that direction.
Yes, they are still around, but they are not employed to the degree they were before the invention of the camera. The same happened with recorded audio and live musicians. It's disruptive technology. People still do it, but it's not their day job.
Very tough business to make a living in now, though.
Nope. He was upfront and honest before during and after the competition about his methods. This has been made clear in every article about this, as well as in this thread itself. People keep missing this point and it's irritating.
"Allen said that he made clear that his artwork was generated with Midjourney, an online AI art tool, when he was dropping off his artwork and in his narrative submission.
Olga Robak, the director of communications with the Colorado Department of Agriculture, confirmed that Allen mentioned Midjourney in his submission statement.
According to fair rules, anybody can file a grievance against submitted items — but they will need to post a $300 bond, cite specific rules that have been broken and present a grievance letter in person.
The bond can be returned if the grievance leads to a violation of the rules but Robak said a preliminary review showed Allen had not broken any."
Presumably the rules of the marathon say somewhere that you have to compete on foot.
If they don't, then yeah you should bring a car or a plane or a horse.
If you care about this sort of thing, just put 'no AI powered tools' in your art contest rules in the future and you will be good to go.
Similar arguments were made about 100 years ago when photography entered the Art world.
This definitely needs its own competition but I'm not in a position to suggest it's not fair or how to regulate it.
With art, maybe artisans will get more work since people may want their picture to be made into reality (mostly with photorealistic objects that were generated).
So when they say it's art. Or it's good or bad art. It really doesn't mean anything at all.
Artists are pretty special in their ability to understand and appreciate (and do) art. They're like a different species.
Have you had a dream where it felt as if it was real? Do you feel the earth guide you to something that feels right? Have you felt a portal of some distant reality exist alongside of us?
Art has been a portal to channel my indigenous ancestors, where I slip under an emotional spinning vortex of creation. The makers of my blood flow through me. I channel the artisans, craft-makers, mud-dwellers, star-makers, dream-weavers and earth-brothers and sisters — the ones who paved the way and forged the path. My work carries spirit and my truth is in everything I create.
The whole technology (especially AI technology, Deep Learning) is a dystopian creation and nothing good comes out of it.
As for this art contest, I would just disqualify digital art all together and only allow hard copy, physical oil on canvas paintings. (And no, not a printed copy)
You can't stop this even if you tried.
Who said stop? I said 'detect'.
DeepFakes have detectors, there are GPT-3 detectors already being developed and the same will happen with DALL-E 2 / Stable Diffusion. I won't be surprised to see someone develop such a detector for both of these AI tools as a research project.
Detection of a deepfake, for example a generated photo of a person of something that never actually happened, may be in the real of possibilities, but still difficult.
Detection of AI in an artistic image? Not a chance.
Fair enough, I'd agree with that - but I think there is no-one left but "art consumers". One could imagine a version of instagram, where the user could scroll through pages of images, all of which were generated by AI. Is that user of Instagram then an artist if they pick a favourite? If I scroll through real instagram and pick a favourite, am I now an artist?
When you look at a high-end photo that you like, one of perhaps hundreds that you see in a day, do you do a detailed background check on the input?
Explore the "life experience" of the photographer? Try to guess some deeper meaning? Study the technical approach and gear used, in order to come to the ultimate appreciation of how this outcome came to be?
No, you don't. You look at it for 2 secs, go "nice picture" and move on.
That sounds like a commodity to me, not a work of art.
None of the winning piece belonged to the winner!
People miss that the tech is here, and it's not going away. We've open sourced it already. It's burgeoning at a wild speed.
Most people seem to miss that this artist was entirely honest about his use of AI assistance.
Many are missing the point that this was a casual art competition; a state fair. There wasn't a huge prize at stake.
Many are missing the fact that there was undeniably artistry in the methods used, including weeks of work, touch ups with Photoshop, etc. I doubt that one in 100 people could get a work like this from MidJourney even given a week of unlimited research and experimentation.
People don't care about these issues. It's ignorance. Trying to have a fruitful discussion on this stuff when basic facts are left by the wayside is like trying to drink water with a fork.