AI Is A Camera
Artificial intelligence, the term we are currently lumping large language models (LLMs) into, is at least 30 years old, so it's not new. It is here to stay. Generative AI is a prediction engine based on that part of human intelligence that involves matching patterns. Its strength is in doing it at light speed. It's also not intelligent. It doesn't think. Humans do. Artists (musicians, painters, writers, et al.) do the thinking and the feeling. Generative AI, as used by artists, is just another tool. The best analogy I can come up with is that AI is a camera.
This post presents my thinking today on using AI in the arts. I'm also writing it in part to respond to the more salient arguments from the anti-AI activists.
🚧 Content Warning: If you're deeply outraged by artists who use AI, please save yourself ten minutes of aggravation and go listen to some calming music instead.
I use AI as a tool because it's helpful. The fun of learning how to use it keeps me off the streets and out of the karaoke bars. The possibility that I can create art — images, writing, music, video — faster and better excites me. My artistic ideas far exceed my ability to realize them in my lifetime. Why wouldn't I use a tool that helps me do more?
The arguments against AI can be boiled down to four main objections:
- The sources were stolen and source creators weren't compensated
- Using AI to make art is cheating
- Using AI isn't creative
- AI steals jobs
Stolen Sources
To me, the “stolen sources” argument is the strongest one the anti-AI camp makes. However, it has flaws.
It's true that the content creators weren't asked, much less paid, for their work to be used to train the LLMs. It highlights the tension between intellectual property rights versus the value of innovation. In this case, the problems are two-fold: availability and scale.
Availability
Availability is sharing. Art shared with others, whether via a painting in a cave or a million words in a bestselling epic fantasy series, is how humans communicate. And how we branch out with new ideas. Courts have found that a fiction story that used a history book as the launch point for a thriller plot does not owe royalties or even credit to the history book's authors.
Let's say that I see a photo of a frost-covered leaf and am inspired to create a watercolor painting of it. What is the difference between my process and an LLM being trained on what leaves look like so it can do the same? Should I send money to the photographer's estate and the out-of-print magazine because the photo inspired my painting? How do I even find them? How much do I send them? Are the answers different if I sell my painting for a profit?
Scale
Scale is size, but it's also doing a lot of things quickly. I'm slow. Even if I had 30 years of experience as a painter, I'd be slow. I'm slow as a writer, too.
Computers are fast. We invented them to save us from mind-numbing repetition and to speed things up. Consequently, it only takes minutes for an LLM to learn how to create something that humans recognize as a leaf, regardless of style variations. That's where scale comes in. An LLM can be taught to accept vast amounts of data, analyze it, and produce what humans want to see, read, or hear based on anything they can describe. It's up to us humans to choose what we like and what speaks to us.
Stealing
Deepfakes are a problem. They cause us to question facts as documented by older technology. However, like other forms of theft, forgery is an age-old arms race between the person who has something of value and the person who wants it. Instead of banning AI (that ship has sailed), we should be looking for ways to shore up our defenses against persons trying to fool us.
Plagiarism is also a problem, but it, too, is an age-old form of theft. Advances in technology merely make it easier and faster. I am, of course, talking about the Gutenberg printing press and the typewriter. And don't get me started on thesauruses. Again, banning AI won't stop thieves. Instead, let's get better at detecting it.
Using Cameras or AI Is Not Cheating
In the 1800s, artistic painters who made a good living were very likely alarmed by the invention of the camera. I'll bet dinner at a fine restaurant that the arguments against cameras sounded very similar to the arguments against AI today.
At root, it's fear: Fear of change. Of the unknown. Of losing something.
“The best way to do this is the traditional way that took me years to learn and is hard work.”
“I don't understand how it works, so I don't trust it.”
“My best work is worthless if anyone with a camera can create a similar good-enough image with the press of a button.”
Is the camera an artist? No, it isn't. Human creativity, ingenuity, and agency are needed to create a masterpiece painting or an award-winning photograph. AI isn't an artist, either. And LLMs don't sit around writing space opera romances in their spare time. LLMs need humans to dream up the idea and use the tools they have to bring it to life.
Cameras, AI, and Creativity
The creative impulse is strong in humankind. When pleasing patterns, colors, movements, or sounds move us, we want more of them. Creative expression gives us insight into the past, helps us process feelings, and gives life to dreams.
Can AI do that? No, it can't. It's just a tool. It doesn't have the desire to design the perfect table, any more than a table saw decides to build it. Cameras don't choose where, when, and what to take a photo of. AI programs don't care if they reproduce a leaf or a starship. It takes human intelligence to feel the need and figure out how to meet it.
The creative process is so uniquely individual to each person. My combination of strengths and weaknesses, interests, and experience is different from yours. Even if we are lifelong friends, your process is different from mine. Using a tool is not “missing out on the creative process.” It's getting to the fun part faster.
For a deeper exploration of this subject, author, teacher, and all-around interesting person Derek Murphy created a fascinating and thought-provoking presentation and a blog post on authentic creativity. It's well worth reading.
Stop, Thief!
Does using AI steal jobs? Probably. However, most technological advancements result in creative destruction. Ask VHS video rental stores, or buggy-whip makers, or eight-track tape manufacturers. But AI also creates new ones, such as prompt engineers, LLM developers, and cloud-computing specialists. Artists are free to stay with the method that works for them to create art. Or, they can learn how to use new tools to create art. Either way, it's still art.
Did OpenAI, Anthropic, MidJourney, etc. scrape the internet for content to train their LLMs? Yes, they did. Do humans do the same thing? Also, yes. In both cases, did the content originators get compensated for this? No, they did not.
No art is created without input. If nothing else, from our senses. The feel of soft fur. The smell of baking bread. The songs of whales. The colors of the Northern Lights. More often, we are inspired by the creative spark in others.
The first non-children's stories I ever read were science fiction (blame my parents, who wanted to distract me one summer). When I got a little older, I added romances to my plate. Without a doubt, the thousands of books I have read in my lifetime influenced my own stories, from the first inkling of an idea, to character choices, to writing style, and even to how the words look on the page.
However, all that pales in comparison to the main reason I write: I want to read that story!
Sadly, my cats are too busy creating mayhem, so the writing part falls to me. If using AI helps me get the story out of my head and into words that I can then shape and sharpen as best I can, sign me up.
AI Is A Camera
I've been saying AI is a camera because they're both tools that enable people with ideas to bring them to life. AI is unflaggingly cheerful and patient. It is never exasperated, bored, or frustrated when we're struggling to describe what we're after. Or just trying to figure out what we want in the first place.
As an art appreciator, I'm easy to please. I want a well-written story that takes me away. I love that painting that makes me smile each time I see it. And I need that music that motivates me to move.
It's none of my business what tools the artist used. I care about the finished art.
