Many Thoughts on AI Art

Art doesn’t exist in a vacuum. If you write a poem and immediately burn it without ever showing it to anyone else, you’ll still have been in conversation with every poem and other piece of art you’ve ever encountered just by writing it in the first place.

Art is the language humans use to communicate feelings and experiences and beauty. Art is fundamentally about connection, whether it’s between someone who creates art and a separate person who experiences it… or even if it’s just a way to communicate more deeply with yourself.

A flower is beautiful if no one looks at it. But it’s not art until someone takes a photograph or writes a poem about it or paints it, because the choices made in how it’s framed and displayed say something about how that artist experienced it and what it made them feel.

Art is fundamentally about connection.

And that’s why I love AI art.

You can call AI art greedy tech bros stealing from starving artists, but at the deepest level, it’s tons of people working together to teach a computer how to synthesize the communal output of humanity’s art into a program that talks back to people in the language we’ve made.

When an AI art program takes a few words and spits out an image, you’re talking to the mass output of humanity’s ongoing conversation in the form of visual art. And that’s amazing. It’s not just a gimmick or toy. It’s a powerful synthesis of what we’ve all made together.

Art styles evolve over time as artists react to what they see in the world and what they see in each other’s art. It’s all a big conversation, and you can’t own a way of having a conversation. You can’t own a style. It defeats the whole purpose of it all.

You can be the first person who thinks to draw a dog a certain way, but if it’s good, other people will start drawing similar dogs. Giving credit to artists is vitally important, but we’ve never been able to own ideas or styles. Just specific creations.

Copyright is designed to let artists control their creations for a little while, in the hopes that it will let them support themselves. I don’t think that’s working very well, but that’s a whole thing. And regardless, it was never meant to last forever.

We already have a massive public domain of all kinds of art, and that’s an amazing thing we’ve all created together.

I’ve been using public domain image archives to cobble together art for my stories for years, and I’ve watched those pages grow and grow in just the last decade.

There are clearly people who go out of their way to create works of art so they can upload them to public domain archives. That’s amazing and wonderful.

And it means AI art programs were always inevitable, because as time passes, there’s just more and more they can draw from.

It’s clearly really important to a lot of artists to be able to opt out of AI art algorithm training sets, and Stable Diffusion, as I understand it, is moving towards incorporating that.

But it’s not going to stop these programs. It might slow them down, but that’s all.

Humanity has been talking to each other in the form of art for centuries, and with computers and digital records, that conversation just goes faster and faster. Maybe an AI art algorithm trained only on public domain works would feel a little out of date in its language…

But that wouldn’t last. Some people will want to opt into sharing their art with AI algorithms, because it’s a way to be part of that piece of the conversation. You can be mad about that, but you’re not going to stop it.

Young artists will grow up using these tools.

I just want artists to remember when they’re raging with fury over AI art and throwing around the word “thief” carelessly…

It’s not just the “greedy tech bros” who hear you. It’s also the young artists delighted by what they’re able to make using AI as a starting point.

And when you’re throwing around the word “soulless,” because you think AI can’t have a soul, try to remember where all the art it’s synthesized came from.

It came from people talking to each other. And when you’re playing with these programs, it feels like talking with them.

Sure, questions of ownership and money and copyright play into all of this, but at a really deep level, typing words into an AI art program and seeing what it gives back to you is a way to connect with the growing communal body of art that humanity continues to make.

Because AI art is human art. It’s all human art. Another creature, like a cat or dog, doesn’t look at it and get something out of it. Humans do. It was made by and for us. It’s part of how we can talk to each other. Not every piece means something. But neither does every flower.

AI art lowers the barrier for entry into making art. That doesn’t mean people won’t still want to learn how to make art in more complicated ways; it just means it’s easier to get started, and once you have someone excited about what they can make, they’ll want to learn more.

If you want a kid to learn how to read, you don’t hand them a thick novel and then sniff about how they must not be serious about reading if they don’t learn how to read it. You start with easy books and work up. Not everyone will become a lifelong reader, but more do that way.

The same will be true of AI art. Some kids have such a strong vision in their heads that even though their hands’ ability to express it lags far behind, they’ll keep working to learn how to draw. Others will give up if you don’t get them excited with an easier entry.

Letting a kid feel powerful by typing words into a program and summoning something like their visions will just make a lot of them hungry for learning how to make their visions get expressed even more precisely.

We’ll have more artists because AI art lowers the entry threshold.

And sure, some kids who play with AI art will never move past that, but that’s okay. Some readers never read long novels; they only like graphic novels. Some people only read by listening to audio books.

The more ways there are to connect to art and create art, the better.

It’s hard to have the courage to write about how truly neat AI art is when there’s so clearly so much hate ready and waiting to jump on you for it…

But I mean, wow, a year ago, computers couldn’t do that. And now, look at what humanity has made.

It’s amazing.

I remember watching Star Trek: DS9 & Voyager as a kid and being surprised and disappointed when they started referring to holodeck programs as holo-novels that had actual authors.

In TNG, it had always seemed like the computer just generated them, and that was kind of magical.

AI art isn’t going away. Stable Diffusion was (as I understand it) made legally in Germany and is already open source.

Corporations that already weren’t paying artists don’t care if you rail against it. But there are small time artists who will feel alienated.

There are valuable ways to incorporate AI art algorithms into the process of making art. And even if all the raw, unedited images produced by the algorithms go straight into the public domain, there are still really cool uses for public domain art.

Should the training and release of these AI art algorithms have been handled differently? Seems like it. But with all the public domain art already out there, that wasn’t going to stop them from happening, only maybe slow them down a little.

So, you can dislike AI art all you want and be mad about how it’s been handled, but the genie isn’t going back in the lamp.

It was possible to bully NFTs away, because they’re genuinely bad and pointless. So, it wasn’t really the bullying that did it.

Whereas AI art does serve useful purposes, whether you like it or not. So, bullying people about it just adds up to bullying, and that’s ugly.

Leave a Reply

Your email address will not be published. Required fields are marked *