People Keep Asking for My Thoughts on AI
June 2023: A letter for people who are, for whatever reason, curious what I think.
Dear family, friends, and Internet strangers,
Hello! I hope you’re having a great June thus far. Since we moved to Indiana, I’ve come to associate June with fireflies, and with open windows at night, except I haven’t wanted to inhale burning Canadian forests and thus have been keeping our windows closed. I’m grateful it finally rained this week, and our air quality has improved.
For some new subscribers who have less internet experience, I want to direct your attention to inline links, something that various online publications have been doing for years now. The words should appear underlined when I’ve linked a relevant article or website to some text. Rarely do I say “click here!” I expect people to notice that there is a link and click if they’re interested. All links should open in a new tab. You don’t have to click anything if you don’t feel like it. You can think of them as newspaper clippings I’ve included in the envelope with my letter.
Thoughts on AI in Visual Art and Writing
Presumably because I’m a professional writer and editor, and an amateur visual artist, people keep asking for my thoughts on AI with everything that’s been happening with OpenAI, ChatGPT, AI-generated images mimicking art, and the like. This month, I’ve finally started to get my face out of the sand and look at it, just a bit.
If you’re not interested, feel free to skip to “Writing Updates” below.
I began working on a complex digital art piece (drawing in GIMP via a Wacom Intuos 4) for our church’s annual summer gallery, and in the process I looked for inspiration and digital painting tutorials on Pinterest. I noticed the comments on some beautiful digital images declaring them AI-generated, like a new scarlet letter.
(Side note: The way Pinterest users go out of their way to mark images as AI-generated is fascinating to me, from a cultural standpoint. While many writers are making the argument that AI-generated images are a valid form of art, the users of Pinterest have collectively said, “No.”)
I clicked on one such piece, and discovered Midjourney, one of the AI programs people can use to generate such images. Specifically, I ended up right on the community showcase page, which shows off user-prompted AI-generated images that look really, really good.
After spending hours digitally painting my own art piece, which still seemed so far from my original vision despite my efforts, I felt devastated. “What’s the point of putting so much work into a piece of art,” I thought, “if a computer can do this better, faster?”
Two threads of thought came from this:
The logical, practical side of me remembered that the algorithms didn’t generate their own visual style. They’ve taken these styles from actual artists. You can see on the Midjourney Community Showcase page (the content of which changes frequently, so I can’t point to specific images for this), as you hover over some of the pieces, that many of the prompts contain the phrase “in the style of” or “by” and then someone’s name.
This situation led to a lawsuit by several visual artists against Midjourney, Stability AI (who develop Stable Diffusion, another AI image-generating program), and DeviantArt. The defendant argues that using artists’ works this way is “fair use” but that’s hard to believe upon viewing the results.
This is no frivolous lawsuit. If you want to keep up with this ongoing case, you can find the updates here.
The second thread of thought I had is related logically to the first, but it’s the emotional or philosophical outcome: Why do we do art, and why do we value art?
Art in America touches on both of these threads in this detailed article about the lawsuit and the issues associated with it: “Artists Are Suing Artificial Intelligence Companies and the Lawsuit Could Upend Legal Precedents Around Art”, Art in America, 5 May 2023
Among others, they quote artist Tomer Hanuka, whose dynamic, modern, and at times unsettling art style has made it onto lots of magazine covers. He describes the encroaching of AI images on his place in the art market:
“It’s passable, it sells. It doesn’t just replace you but it also muddies what you’re trying to do, which is to make art, find beauty. It’s really the opposite of that.”
Hanuka has raised multiple discussion points here.
An AI-generated image can be beautiful, in the visual sense. Is that enough, or do we mean something else when we seek “beauty” in the arts? Do we value the meaning we perceive in an image, or the creator’s intended meaning? Must art have meaning to carry beauty?
That is, does it matter whether a human created it and intended anything for it? Or do we just like when it’s pretty?
I appreciate aesthetic beauty for its own sake. Yet I find looking at AI-generated images, knowing they’re computer-generated, destroys much of their value to me. Somehow I have this idea that a person pouring his or her soul into artwork creates some large percent of its value.
I also notice that Hanuka used the word “replace”, as if AI-generated images can actually replace his art. For some, they can, if the goal was to include a certain style of illustration. But illustrations in the style of Tomer Hanuka could not exist without his entirely human input, which he didn’t give by choice.
If there are “logical” arguments against the lawsuit, such as fair use, the fact that a style cannot be copyrighted, and the like, it is also well within reason that something is skewed if people can make money off Hanuka’s hard work without Hanuka’s permission, and push a human artist out of his own niche in the market.
Finally, he brings up the age-old question, “What is art?” by casually stating that AI-generated images are, in fact, the opposite. I don’t know what the opposite of art is, but I am, for the moment, in agreement. (In fact, he’s not the only person to declare this the “opposite” of art; for a more detailed look at the topic, see “ ‘It’s the opposite of art’: why illustrators are furious about AI”, The Guardian, 23 January 2023.)
The court of public opinion may decide this before the actual courts do, if Pinterest comments are anything to go by. One user commented on an allegedly AI-generated image, “The style is stolen.”
Two points worth noting:
First, actual artists steal from other artists, in a way; we copy, or draw inspiration from, the styles and techniques of others.
“Nobody is born with a style or a voice. We don't come out of the womb knowing who we are. In the beginning, we learn by pretending to be our heroes. We learn by copying.” -Austin Kleon, Steal Like an Artist
As I mentioned above, I’ve been seeking inspiration on Pinterest for my own drawing, and I’ve been actively seeking to understand what I love about various artists so I can incorporate such into my style. But the goal is to generate my style, a human affected by humanity, not to replace another person.
Kleon created a useful set of guidelines contrasting “good” and “bad” theft, helping us to sort out what the difference is. He includes contrasts like honoring vs. degrading the original artists and their works, studying vs. skimming, stealing from many vs. stealing from one, and crediting the original vs. plagiarizing. If we agree with his premise, the difference between my “theft” in this example and AI’s “theft” in the example of Hanuka’s “replacements” becomes clearer. But it does not appear that all AI-generated images land neatly in the column on the right.
There are uses of AI-generated images I wouldn’t necessarily balk at. If computers can generate, say, backgrounds for webtoons by artists who already take shortcuts in that area, because they focus their attention on characters and objects in the foreground, or illustrations for webpages that just need any relevant image, or other situations where the desired image is only a soulless product, without stealing substantially from other artists—stealing like an artist—I think there’s plenty of room for that in the marketplace.
But I think the soul of the maker is the point, when it comes to most art and writing. We seek art and we do art because we were created by the creative God. At their best, our creations reflect the God who made us. They are greater than the sum of face-value content and aesthetic.
The other point worth adding here is the rules humans make don’t always have the intended effects, as Cory Doctorow pointed out a few months ago: “Copyright won't solve creators' Generative AI problem”, Pluralistic.net, 9 February 2023
Doctorow has different views on freedom of access for created content than most of us, even offering his own books for free. I appreciate the perspective his writeup offers and the examples of unintended consequences in creative industries. He could be completely correct that the aforementioned lawsuit is not the solution to this problem, and depending how it’s handled, could make more problems than it solves.
As for whether I’m worried about my writing being replaced by robots, no, not right now. ChatGPT generates false information, misunderstands information, and cannot understand the market or purpose of a piece of writing, besides other problems. When my best market financially is technical white papers, which so far require a level of understanding these chatbots haven’t achieved, and my desired market is fiction, which requires a level of humanity to be done well, I don’t see the likes of ChatGPT replacing me, at least not soon. (I may eat these words down the road.)
However, it sounds like content writers who were generating blog posts trying to keep up with SEO (search engine optimization) are, in fact, being replaced, at least in part, with AI. My understanding is that those kinds of posts were being made for robots, not really for humans. I hope those writers can find other work. I also hope this means we’ll see less of that kind of content, not more; I hope, with blatant optimism, that the absurdity of robots writing webpages for robots brings an end to such.
This post is narrow in scope and doesn’t begin to cover things like the potential innovation available to, say, NASA, or the pitfalls of trying to use AI tools as a work shortcut (e.g., accidentally revealing trade secrets).
This post doesn’t cover the bigger, scarier issues of what AI could theoretically do as it continues to be developed; I really don’t have enough information to make any comment on it. As long as other, more authoritarian nations are working on AI for nefarious purposes, it makes little sense to me to try to “ban AI,” whatever that means practically, in a nation learning to use and improve them for other purposes.
Let’s find the good uses for these tools and create boundaries, cultural or otherwise, as we need to, basing our choices on better things—not greed, nor wild fears. May God help us to move forward with wisdom, truth, and grace.
Writing Updates
This month, I was focused mainly on finishing my art piece for the summer gallery show. I worked a little on my novel, and had a great time meeting with other local writers in our little Inklings group.
The short story I wrote last month was rejected. The feedback: It’s a good story that’s too similar to others they received. I am planning to rework it eventually and submit it someplace else.
For now, I want to focus on my book alone—until January—and of course continue this newsletter and my paid work.
All my writing is still 100 percent organic, no AI content added.
5 Pillars of Dementia Prevention
June is Alzheimer's & Brain Awareness Month. Last year, Ed and I attended a Star Trek event with some friends, an interview with Kate Mulgrew of Star Trek: Voyager and now Star Trek: Prodigy. The money went to benefit Alzheimer’s research, and because of that, a person from some Alzheimer’s-related group or foundation (I took zero notes on that detail) gave these five pillars of prevention:
Sleep: Most Americans do not sleep enough!
Nutrition
Exercise
Continuous learning
Life purpose and community
Some of this (2, 3, and 5) overlaps with research from the Blue Zones, the places in the world where people most frequently live the longest, healthiest lives. I think the places were originally circled in blue ink on a map, hence the name. I want all my friends and family to live long, healthy lives, so do your part to reduce your risk!
Random
I am persuaded that the most important phrase to learn in any language is “thank you.” This isn’t simply a profundity; it’s based on my habit when trying to communicate in any language at all. Gratitude for all the grace given to me when I do not speak someone’s language can be enough, for the moment.
That’s it for June! Substacks grow best organically, so if any of this was valuable to you—interesting, useful, or beautiful—share it with someone who’ll appreciate it:
To truth, love, and adventure,
Rae