I recently read an article on Ars Technica titled, “From toy to tool: DALL-E 3 is a wake-up call for visual artists—and the rest of us.“
tl;dr: The article talks about a new AI tool called DALL-E3 that has the unprecedented ability to create any kind of art in any style.
It piqued my interest because I’ve been thinking a lot about AI-generated art and writing over the past year.
First, I want to make clear that I’m not opposed to the OpenAI project at all. As a computer science nut and SysAdmin, AI fascinates me and I think it will be a useful tool in lots of areas for years to come. It’s fun to develop AI scripts. It’s also fun to mess around with ChatGPT and see what it can do.
But in terms of creating art, I have a different take.
On the writing side, I had a client who asked if I would be open to using ChatGPT to generate articles, then tweak them afterwards. I immediately said no. Two main reasons drove my response:
- Because ChatGPT is pulling from other articles on the web, I believe it’s a form of plagiarism when used for publishing content under the guise that it’s an original work.
- I think it’s lazy. Writing is hard work, and I’ve become good enough over the years to get paid for it. Using AI to generate articles I would get paid for feels like cheating.
Some people see it as merely a tool to make doing hard work easier, like using a computer to calculate large numbers quickly. I asked myself if I were a mathematician, would I refuse to use a computer? Of course I wouldn’t. But the origination of mathematical theorems needs to come from a human first, not the other way around.
Animators are famous for being lazy and finding any kind of trick to make the process less work. As an animator, I agree with that idea. That’s why studios like Pixar use 3D animation tools to produce the animation instead of drawing each frame by hand. However, the animator is still on the hook for making the characters move in a believable way, as well as express a multitude of emotions. That’s called acting, and the computer can’t do that.
What about art? The article makes a case for the idea that AI is getting better at “executing ideas,” but what it means is that it’s more capable of executing human ideas, not it’s own. AI doesn’t decide on a whim to generate an image of a robot bunny warrior in the style of Boticelli. A human needs to tell it to do that, of course. But what about the rendering of the art itself?
The AI isn’t pulling out specific images from the web. Since Botticelli never painted robot bunny warriors — that we know of — it’s a good bet that the image doesn’t already exist. It’s pretty specific. Looking back to my plagiaristic writing concern, I don’t believe that AI is directly stealing some artist’s work. It’s representing the painting style of Botticelli, but not the actual content.
One concern about AI-generated illustrations is that publishers will no longer need to hire an artist to depict an image for an article or a book cover. All they have to do is ask DALL-E3 for (literally) whatever they want and poof! Done. Some commercial artists will likely suffer financially as the world starts using this tool on a regular basis.
However, this was already happening without AI. As a professional illustrator, I often found people who wanted to hire me because they loved my work, but I was too expensive. So they would hire a much cheaper artist to create something in my style. I once got an email from one of these artists, asking me to instruct her on how to produce the kind of artwork I made so she could complete the job I wasn’t hired for. No, I’m not kidding, and yes, I declined to give away my method.
So the problem for artists isn’t DALL-E3 itself, but the people who choose to wield it. The inclination to have art produced cheaply or for free already exists.
As an artist, fighting to compete with AI is an impossible task. We can’t be faster and we can’t survive on electricity. Beyond wringing our hands and worrying that we’re all doomed to become slaves to robots in a dystopian world, what can we do?
To put it simply, we can be human.
I believe there’s still a need and a market for human-made art. There’s something about standing in front of an original painting or drawing and marveling at the skill. It’s fun to think about the choices the artist made while they were standing in front of the canvas or paper when it was blank. Why is that brush stroke bolder? How did they create that perfect reflection of light? What did they listen to while they were creating this piece? Is that blurry spot a happy accident or was it done on purpose? Humans love to own something that another human made with their own hands. That has value and will never change.
Sure, someone can use ChatGPT to generate a novel that reads kind of like Stephen King. But that’s not the same as reading a novel that Stephen King thought up himself, is it? There’s no romance in thinking about a chip spitting out bits of information that assemble into a story. And reading, if you love to read, is a romantic activity. A human activity.
As artists and writers, it’s our job to create works that feel human. They’re messy in places, maybe the grammar isn’t perfect, there’s a spot of ink in the bottom left corner. That’s real art. Something that was worked on by a human.
I believe people will still buy those things. They’ll seek them out. And some people won’t. Some people will always go the cheap, lazy route. That’s not the right market for your art.
Keep making your kind of art, your way. The world needs it.