You must log in or register to comment.
What I hate about AI art: How it’s based on stolen work. How it is purpose built to replace real, talented artists and devalue their labor. How it uses way more energy than it needs to and is pretty wasteful
What I love about AI art: Instant stupid shit for meme madness.
If AI art was all just stupid jokey shit like this that a friend of mine made when we were discussing how people were making Ghibli-fied versions of important moments in history, and we decided to go with “George Bush doesn’t care about black people” but make Mike Myers dressed as Austin Powers, I’d be okay with it entirely. It’s not for profit by devaluing artists and using this work instead of a real artists work, it’s just stupid shit that makes us laugh. Everything else aside, I can get behind stupid shit that makes us laugh. The rest of the issues with AI art suck though.
I’m with you on this one. I have no issues with AI being used for shit posting and memes, other than the ecological impact I guess.
In general - yes. There is a flood of shitty and lazy “art” that has infected search results and creative spaces. I’m also deeply uncomfortable with it being trained on artists work without their consent - for all the talk about it being equivalent to human inspiration I’m pretty sure there have been examples where it’s started generating attempts at signatures.
It’s terrible in knitting and crochet spaces (I imagine woodworking and sculpture and architecture too) because there are lots of things generated which are physical impossible and just wrong to anyone who enjoys the crafts. It gives false understandings of what those art forms look like.
I think the entire point of art is the human intentionality aspect. Art is humans using materials to do things that don’t serve an immediate practical purpose. There has to be some element of “desire” on the part of the artist.
So it’s not that it is impossible to use AI tools to generate art (there’s stochastic computer generated pieces from the 70s that are lovely iirc) To me though, the way these tools are used is what is important - if you’re using an AI you’re training and adjusting yourself, if you’re spending hours tweaking prompts and perhaps sifting through hundreds of pictures to combine and really participate in “making” something.
The current trend is really just a bunch of content sludge. I don’t see the appeal in either the process of creation or in what can be appreciated from it. The best stuff is mostly memey topical political jokes, where it rests more on the symbols rather than the art itself.
Like, when I make art - my process is adding layers over weeks and weeks. It’s noticing that I don’t like the way this section looks, so I go back over it, come back to it later… it’s a process - I engage with and shape the work. I’m just a guy who glues trash to things and paints them, my art doesn’t really have external value - but it still feels like art in a way that getting Midjourney to make pictures of Gandolf with big honking naturals isn’t.
I don’t hate AI art. I hate people who pretend they’re artists when all they do is writing prompts.
It’s soulless. A mere imitation.
Yes, I hate it. I hate that it fills every image platform. It is not art at all.
It’s a fun toy thing and can make decent images but its not art and can never replace actual art. When you compare for example an anime art of someone who actually drew it and the AI image, the drawn art is 9 out of 10 times better.
It’s also petty pretty easy to spot whether an image is AI or drawn made.
It’s also petty pretty easy to spot whether an image is AI or drawn made.
Doubt. Most studies have shown that people are horrible at actually picking out AI art. You suffer from selection bias because you don’t realise which ones you didn’t spot.
its not art and can never replace actual art. When you compare for example an anime art of someone who actually drew it and the AI image, the drawn art is 9 out of 10 times better.
That implies it’s solely about quality? At the inevitable point where AI gen gets better than drawn art, is the AI gen image now art too?
I’m not entirely against LLMs as a tool, but I especially despise the image-based LLMs. They are certainly neat for some fun things. I’ve used them a little bit here and there for a dumb profile picture or a “I’m kinda thinking about this…” Brainstorm, but even in those cases I noticed the capabilities of the LLM and its tendencies quite literally pidgeon hole my artistic vision and push me in other directions that felt less and less creative. (Sidenote: I feel the same way about coding LLM tools. The longer I use them at any given time, the less creative I feel and it has a noticeable impact on my interest in the code I’m writing. So I don’t really use them much. Also I consistently manage to point out coding LLM code in PR reviews because it’s always kinda funky)
I’ve avoided using AI art tools for a while now. I’ll consider some limited use if the cost, billionaire ownership, blatant theft of real IP without compensation, and environmental impact problems are solved. (No, an “open source” model doesn’t solve all of these problems, especially since nearly all open source models are not truly open source and are almost always benefiting from upstream theft)
You know what I do like about AI art? I like the older Google machine learning art experiments from the mid-2010s. They invoked a strange existential curiosity. But those weren’t done with LLM’s.
Outside of LLMs, I like that there are some newer tools for editing that can do a better “lasso” select, that can mix and match into brushes as an alternative to something more algorithmic, the audio plugin that uses a RNN to simplify or expand upon an audio technique. Things that are tools that can be chosen or avoided and have nothing to do with LLMs.
I honestly cannot wait for this bubble to burst and for these tools to return to a cost that they’d need to be for these companies to turn a profit. A higher cost would eliminate all this casual use that is making people worse at research, critical thinking, and creativity, as well as make the art tools less competitive to just paying artists, even for scumbags wanting to cut the artists out. And it’d incentivize non-LLM, non-insanely costly ML techniques again instead of the current “LLMs for everything” nonsense right now.
It’s pleasant to see such a long post so full of good takes. Nice work!
I’m not sure hate is the right word. When you’ve got someone stabbing you in the back multiple times, is it really hate you’re feeling toward them? Or is it anger, fear, and danger?
I “hate” it in the sense that it’s built on theft and requires the exploitation of underpaid workers to develop and maintain it. I “hate” it in the sense that we’re living on a burning cinder with dwindling fresh water resources and “AI” is adding fuel to the fire. I “hate” it in the sense that it’s being used to further undervalue artists and writers. I “hate” it in the sense that it fills our spaces with crap that so often looks like it was cribbed off of Rapunzel, Wreck-It-Ralph, and some other things.
I don’t hate the “art.” The AI can’t do much about it.
What I strongly dislike is people who manage to draft literally 40 words or less and think they “created” something.
You didn’t. You a mathematical model to do something for you. You therw 175 tokens into a whirlpool and got am 87% what you wanted image out. If you even had an idea of what you wanted before hand.
No, because I don’t have an irrational fear of AI. I don’t like when poor or unfitting AI art is used, but it isn’t AI who makes that decision to use it.
It’s ruined art for me. Someone posts something, and I don’t know if it’s real art or a theft of other people’s work.
That’s the problem. We can’t tell the difference.
Art is cool cos it’s like holy shit a person did that!?
If it’s just an algorithm it’s not very impressive.
I hate those who call themselves artists when they’re just commissioning a computer to make a picture for them. I also hate it when those same people deny the unethical aspects of AI generation.
Edit: to add more, I also hate the AI images themselves. They are filling up the internet with slop. This is very annoying, and the same goes for LLMs. I don’t want to get AI generated results when I didn’t search for them specifically.
Art is about expressing one emotion from one person to another.
We have a word for fake pictures: advertising.
The first phrase is true.
The second, I’m not sure. Some really talented artists have worked in advertisements for a long time, and many of their works are celebrated internationally. Alphonse Mucha is one name that quickly comes to mind - tell me his advertisement work isn’t art. You have probably seen more amateur ripoffs of his style in your life than the real deal.
Yes, as it conveys nothing more than the prompt it was given. Art is a means of communication, but when all it does is chop up pictures it’s seen to match a prompt there just isn’t anything to analyze.
It may look pretty in the moment, but lacks all substance and will be forgotten as quickly as it was generated.
Just playing devil’s advocate here. Let me lay out some counter points … (it’ll take me an edit or two to format this right, btw.)
-
Instructing a machine to assemble bits in a specific way takes creativity. My prompt to AI is that creativity and without it, you can’t even get much of a copy of anything. Even though AI is generally assembling stolen bits, the end result (ignoring copyright law) can be original.
-
Music has been mostly “figured out” and many songs we have heard over your lifetime use many of the same exact chord progressions. I-V-vi-IV being one of the most common and used in the following songs:
Journey – “Don’t Stop Believing”
James Blunt – “You’re Beautiful”
Black Eyed Peas – “Where Is the Love”
Alphaville – “Forever Young”
Jason Mraz – “I’m Yours”
Train – “Hey Soul Sister”
The Calling – “Wherever You Will Go”
Elton John – “Can You Feel The Love Tonight” (from The Lion King)
- Musicians may use patterns or progressions from other songs. Painters may use the same colors and brushes designed by other artists. In both cases, techniques that have been known for thousands of years are being used in self-expression.
I assert that given the correct instructions, you can still give someone plenty to analyze, via prompt, that has enough detail to extract a deeper meaning:
FWIW, I am extremely fed up with this AI hype now. “AI” is just a tool, and that is it. I could go on for hours about this mess, but I am trying to make a valid point: Regardless of how you interpret copyright, art is just self-expression.
There are endless examples I could give about technique re-use when it comes to creating art with machines. From my perspective, a particular brush stroke might be the same as using a specific bit at a particular depth of cut on a CNC. The art theft for AI training is one aspect, for sure. The biggest issue I see is that many people don’t understand how to create original art and the AI just spits out a copy of something it was trained on and something the user already saw.
Edit: After reading many of the other comments here, many people have a strange definition of “art”. Yes, art can be about communication, it can be about sending a message, it can express a style of creativity or hundreds of other things.
Art is just… art. It’s something a person sketches, composes, speaks, signs or farts. You don’t have to like it or agree with it. Hell, you don’t even need to recognize something as art for it to be art. Art is just self-expression. It’s a feeling that is converted into some kind of other medium that others might happen to see, feel or hear, smell, taste or a combination of all of those things.
As much as I hate to admit it, a banana taped to a wall is art. Someone eating said banana is also art. I think it’s fucking stupid, but who am I to not call it someone’s self-expression?
That some, most or all art is partly or wholly derivative of other art is not relevant because the process used by ‘AI’ does not resemble the artistic process. When Shakespeare wrote Hamlet (a work derived from an older play, itself derived from an older myth which itself had been through countless retellings, variations and translations), he did not do what an LLM does, which is approximately to say: ‘It’s statistically likely that the phrase “to be” will be followed by the phrase “or not to be”’. Putting together statistical likelihoods is not creativity. This alone shows that AI ‘art’ is not creative and therefore not art at all.
Additionally, instructing a machine to make things from prompts does not require creativity. Creativity is not ‘having ideas’; it’s an ongoing process. When you tell an image generator to make an image, you’re not asking it to create something, because it cannot do it. You’re saying ‘Show me the statistically likely output for this input’. Again, this statistical generator is not the same as, nor is it comparable to, the human imaginative process.
Example: My picture of the Lemming.
I knew exactly what I wanted before I even typed in the prompt: My vision was for a nervous, burned-out lemming to be sitting on a log, hunched over a laptop smoking a cigarette with bloodshot eyes surrounded in crushed beer cans.
That is not creative? Saying I have no imagination or creativity is kinda rude and giving all my credit to AI is downright insulting. Sure, I didn’t draw it and I absolutely do not have the ability to draw it. However, you cannot (reasonably) deny that the idea is mine. I’m not exactly the most creative person in the world, but damn… (The image will show up under my username over at least two instances over the span of 1-2 years? It’s mine, is my point.)
If you saw my edit, you should know exactly what I thought when you said “artistic process”.
However, my underlying point about derivative process or technique was to shoot a hole in the arguments of “cobbled together bits from wherever” and why I specifically used music as an example. Drum lines are openly copied. Not derived: blatantly copied. It’s considered a compliment in many cases, actually. Progressions and transitions are all just copies. You don’t even need AI to “statistically generate” music patterns. With every chord I choose to start a progression, there are only X number of chords that will work correctly after it.
I believe there have been some projects to generate (within reason) every chord progression possible and every kind of melody that would fit it… statistically. Almost every bit of popular music you hear is a derivative or a copy or reused or whatever, is my point. How many times have you heard the “Amen break”? More times than you actually know, unless you know your music, then you do. Much of music is just, for lack of a better term, math.
Creativity is an idea or multiple ideas. It’s anything that exceeds the sum of your existing knowledge. AI by itself isn’t “creative” and it is impossible for AI to be creative, we both agree. Again, from my perspective, AI can be used as a tool to fill in the gaps between two different ideas. It’s the assembly of different ideas or components that is important. The sum of the key bits.
In my CAD work, I use formulas and simulated physics to automatically generate connecting features or structures. Are the designs I create exempted from “art” because of that?
Putting creativity and art into a box and saying you must follow “creative process” or “artistic process” is just odd. You can think that way if you want, but it’s very limiting. The artists I study make a habit of saying “fuck the rules, fuck the process and do what makes you feel good.”
Just for lulz, I was wondering what another machine would think of my Lemming. It kinda got it, but kinda didn’t. Statistically, it figured out the parts, but you should know darn well what my intent was:
Again, you’ve written quite a long comment, almost none of which is pertinent.
Music is not math. Some aspects of it can be expressed mathematically, yes, but that’s not the same thing.
Imagining the idea ‘I’d like to see an image of a lemming’, which is what you’ve done, does require some imagination. However, the output is not art because the process used to go from your ‘prompt’ to the image was not a creative one. (Also, this isn’t entirely pertinent, but the image output is really bad. If it had been made by a person and otherwise looked like this, I would still say that it was just ugly, bad art.)
You may well be a creative and imaginative person; I don’t know you and I wouldn’t want to judge! However, your image of a lemming was not the result of a creative process and so is not art.
Sorry if you couldn’t see my points, but that is OK.
I’ll go off and create some corporate logos then…
Having ideas is not creativity. Creativity is creating the thing. If a billionaire pays a painter to create their idea, the artist is still the painter, not the billionaire commissioning the art. Replace the painter with AI and the logic doesn’t change, the person putting the prompt is not an artist. It did not create the thing. The machine is not an artist either, as the human painter at least had consciousness, intention, agency, emotion,all things the machine doesn’t have and cannot source from to create the art. This is why AI images always feel soulless, dry and boring. They don’t produce any emotion on the audience because it had none to source from or communicate through the art. The prompt engineer is no artist but a commissionner to an inept soulless painter.
We are talking about a word with multiple definitions and it’s getting philosophical now. Depending on where you look or what context you use, creativity is how you choose to define it. (I hate saying that here because, well, its philosophical and any back-and-forth rapidly becomes subjective. On the intertubes, that usually doesn’t work out very well for discussion.)
https://en.m.wikipedia.org/wiki/Creativity - Creativity is the ability to form novel and valuable ideas or works using one’s imagination.
https://plato.stanford.edu/entries/creativity/ - We may ask the same question not just of artworks but of any creative product, whether it be a new scientific theory, a technological invention, a philosophical breakthrough, or a novel solution to a mathematical or logical puzzle. (There is more to this regarding creative process, so feel free to read more.)
https://dictionary.apa.org/creativity - the ability to produce or develop original work, theories, techniques, or thoughts. A creative individual typically displays originality, imagination, and expressiveness.
A discussion about a definition is usually fruitless. I just have to cap this saying that I simply maintain that AI can be a tool for creating original art. (Art doesn’t need to be a painting or a picture.) We, as creative humans, can create art with any tool we are given.
If you want to get philosophical, please do. I won’t argue my point further that it still takes a human to provide creative input to get some kind of unique output.
thanks for dropping the definitions here :D
so there’s no thing called “thinking creatively”?
The problem is AI art can never be any deeper than the prompt and can never hold up to anything more than a surface level analysis.
I actually agree with that. I don’t actually appreciate AI generated stuff more than what we see on the surface. It’s not highly complex and I never said it was.
(I was just arguing AI generated stuff can be a means to an end and still carries a hint of creativity as stated by the original prompt. It’s a tool. Personally, I detest nearly all of these LLM parlor tricks. I think people who were giving counter points thought I was pro-AI stuff when I really am not.)
I have a reply to most of the points you’ve brought up which I hope will help you see another perspective. Some things I hadn’t even thought of until you wrote this (thanks). But I don’t have time to write them all now, nor do I want to type it all out at my phone. Leaving this comment as a reminder.
Cool. I just realized how many definitions there are for the word “creativity”, so I posted a reply to another comment that explains which one I was using. It may help frame a response for you, it may not. This is just an attempt to keep the forks in this road to a minimum. ;)
based on your points (2 and 3), do you think there are no more “original art” in this world?
Well, to shorten everything down, there are few original building blocks. “Original art” can absolutely be original but, at the end of the day, is usually made up of large blocks of existing stuff. How those blocks function as a whole could be considered the “original” bit when talking about my specific arguments.
I mentioned music because there really aren’t many methods that haven’t been discovered and used hundreds of times over. If I had a nickel for everytime I sketched out a four-to-the-floor drum variation, I would be a wealthy man. (Nearly every EDM/House/Techno beat uses something like it.) Chord progressions and transitions are best learned from the classics: Beethoven, Mozart, etc… (Hip-hop/Pop artists are notorious for trying to sue each other over sample theft even though the samples are generally common musical phrases… Eesh… That is a mess of a situation, btw.)
But. Original art made with original components can still be made. This is a big world, after all.
To clarify though: I do not support the theft of data for the use of training AI. Period. I also don’t particularly care for AI art, personally.
thanks for your opinion 🙂
-
AI art is fine being used as a tool. What I have a problem with is it’s users calling themselves “artists”.
A person who types a prompt into an AI is no different than a person who hires a painter and describes what he wants them to paint.
Just because that “painter” in the first case happens to be a computer, that doesn’t mean that by default the title of “artist” defaults back to the person who wrote the prompt. That person is still just someone telling someone (or something) what to draw.
In other words, you don’t become the artist just because you eschew paying an actual artist and instead have your computer do it for you.
I see this view a lot and I agree with it, but how can you argue with anyone who says well, that’s what they said about synthesizers for music. Or ipads for art. Or computers replacing typewriters. Are you saying anyone who doesn’t hand write a book in cursive while solely learning the djembe from a scroll is not a real artist ? (Obvious /s but it’s kind of valid). Unfortunately we are the old man yelling at cloud, just 20 years too early. The future will laugh that there were AI detractors (in my opinion).
Everything you’ve mentioned are tools for an artist to use to express THEIR talent. A typewriter doesn’t come up with the words. a Synthesiser doesn’t compose the the music that its playing. Comparinging AI (which requires zero talent) is disingenuous.
To put it another way, if you’re a carpenter using hammers and saws (tools), and then some engineer creates a robot that can be programmed to do that job and allows them to fire all the carpenters. Does that make the programmers carpenters even though not a single one has used a circular saw.
The line between “tool” and “crutch” is drawn by how much talent and training it takes to use it.
AI is NOT used as a tool in that traditional sense, its a shortcut to fake talent in ways that hammers, paintbrushes, typewriters and even just good old fashioned traditional Photoshop aren’t…
You have to have the training and talent to get use out of a real tool. And AI certainly potential for use in that regard; proofreading, background removal, grammar checking etc…
Your comment made me think of DJs. Not “real” musicians? Seems like a similarly-structuree argument.
DJs use the tools to express THEIR talent. They don’t just say “create a composition that sounds good”. That’s the difference between an artist and a fraud.
Are modern DJs musicians as compared to a classically trained cellist (sp?)? In my thoughts, not really.
