When AI image generators, like Midjourney, were finally able to generate semi-adequate images, I started using them in my D&D content. Previously I had relied on creative commons images, purchased clip art and made my own images. When the moral uproar took place over AI in general and image generators in particular, I wondered if my use of AI images was morally wrong. The answer is, of course, yes. As our good dead friend Thomas Aquinas argued, every deliberate human action is either morally good or bad. Using AI images generated by AI is obviously bad. Like all humans, I do many bad things. But as a professional philosopher I can present an analysis of just how bad I am.
As a sometimes-self-aware person who can admit that I do many bad things, I try to explain why I do these bad things—often to re-train myself to stop doing them. What follows is an explanation of why I use such images. I am not offering a justification or an excuse—this is all about why I do what I do with no attempt to excuse or justify it. Also, since we humans are usually garbage at sorting out why we do what we do, what follows is likely to be inaccurate and full of rationalizations. Jokingly, my view of human behavior is that we do what we do and we don’t know why. But we love to lie about it.
One explanation is that I love technology, especially computers. My dad taught math and computer science and when he went to school on the weekends for some work-related stuff, he would sometimes bring me. To keep me from running around like a wild animal, he let me use the DECWriter II at the school to play games. This is how I became a computer gamer. I remember making my own book about robots and computers; I think I called it “the computer pages” and I added all sorts of (not so) special effects, like pullout tapes of paper and things involving tin foil. I continued to use computers, such as the Trash-80 and then got my own Commodore 64, which was (and still is) an amazing machine. I was involved in AI debates in college during the 1980s and ran a Traveller campaign focused on an AI bent on destroying humanity. I went to grad school when the Mac Classic came out and was blown away by MacPaint and MacDraw and my love of computers just got worse. I’m writing this using a AM4 gaming rig I built (and have to keep hanging onto because of AI), and I’m surrounded by a technophobe’s nightmare of modded tech.
When AI began to emerge, I predicted it would kill us all (which it did) but was thrilled to have “new” tech to play with (I know the history of AI). That is part of the reason I use image generators. This might explain why I think AI image generators are cool and like them but one might point out this does not fully explain why I use them in my D&D adventures. This takes us back to MacPaint and MacDraw.
When I first started writing for gaming companies in 1989, I had an electric typewriter and graph paper. I hand drew the maps on graph paper and typed up the adventurers I sent off. Then I bought a Mac Classic. I got MacWrite, MacDraw and MacPaint and immediately started using them for my classes and game writing.
The word processing and graphics software let me do two very important things. The first was that they provided a way to easily keep copies of my work. When I did a map by hand or typed up a Dark Conspiracy adventure on my typewriter, I had one copy that I would need to take to the library and photocopy. But with a computer, I could back it up on disk and print as many copies as I wished. Although disk failure was (and is) a thing.
The second thing is that software allowed me to do edits without erasing, starting entirely over, or using whiteout. I remember that many times when I typed something out and realized I could do better, I usually convinced myself it was good enough—because changing it meant re-typing it. I did drafts; but usually a handwritten outline. I do know that great writers claim they retyped things hundreds of times but as a kid in grad school making a little extra Ramen noodle money, I did not have the time to do that. But with software I could edit and revise text and images easily, so a work would go through many changes for the better. While my focus here is on images, back when word processors came out it was (of course) claimed that they would ruin writing. They did not. But LLMs might.
While I did the rough drafts of my maps and images on paper, I used MacDraw and MacPaint to do the final versions. These tools made map and image creation faster for obvious reasons. Then I saved up enough to buy Aldus Freehand and Adobe Photoshop, which blew MacDraw and MacPaint away.
After getting my PhD. I started teaching at Florida A&M University in 1993. I was assigned to teach Aesthetics, which I dreaded. That had been the domain of the beret wearing crowd who smoked weed and agonized over Proust and Derrida. But I found that I was operating from prejudice (see, I told you I do lots of bad things) and eventually enjoyed teaching the class. During the 1990s one of the “big fights” in aesthetics was over Photoshop and similar programs. Photoshop and digital image creation also launched a moral panic.
One concern was that Photoshop would make it impossible to discern between real and fake images. Of course, anyone familiar with the history of photography knew people were manipulating images since the invention of the camera. While Photoshop did create challenges, these were overcome. The same panic has arisen over AI image generation, and it is worth considering that this time the panic might be justified. I do condemn using AI images for deception, but I use them for my D&D content—my goal is not to trick people into thinking that unicorns are real or that Republicans are working with demons.
There was also the debate over whether digital images could be art and criticisms that they could not be art because Photoshop made it “too easy” to create images. Ignorant critics just assumed Photoshop created images. Informed critics lamented that filters, actions, and such did too much of the work. Defenders pointed out that Photoshop and similar programs still required skill (and new skills). People who used it thinking it would make them automatic artists were disappointed.
Using an analogy, drawing by hand is like getting around by walking or running—it is all on you. Using Photoshop and Freehand was like using a bike—it enhanced what you already had, but you still needed to do the work. As you would guess, I was an early defender of digital art and played a minor role in the Art Wars of the 1990s. While I am sure there are still people who are still mad at Photoshop, digital art tools are the norm today. So, when AI image generation arose, to me it was just another tool, and it faced the same panic as Photoshop. But one might say, surely AI is different?
I won’t say that they say that about every new tech because AI image generation is different in ways that can justify the criticisms and even the panic. Going back to my analogy, if drawing by hand is running and using Photoshop is biking, AI image generation is using a self-driving taxi (which fits especially well since there are also humans behind “self-driving” taxis). Using an AI image generator for D&D content does take some skill but they are not the skills one uses to create art. Which is one of the many reasons that AI images are not art and people who use AI to create images are not artists. This is why I call the “art” in my D&D content “images.” They ain’t art.
In terms of the skills I use when employing an AI image generator, I need the skill of creating the monsters, magic items and such that will be in the images. This might be art but probably not. As I always argue, if I can do it, it ain’t art. Then I need to create appropriate prompts and review the images. This is writing but not art. The final step is loading them into Photoshop and fixing all the AI problems (like extra arms). This is just touching up images, so is a technical skill and not artistic skill. Again, If I can do it…
I can do all this quickly—the longest part is obviously coming up with monster designs, magic items and such. This is one reason I use the image generators—they are fast. As writing D&D is a hobby and not how I get the income I need to buy food and live under a roof I need to use my time wisely. I’m running a 4-hour D&D game every week, so I need to put out material fast—and I like having images for the tokens and to show to the players. This is one reason I do not hand draw my images—the maps and text are essential to running adventures while the images are a luxury. So, I spend my time on the maps and text. But why don’t I hire artists instead of using AI?
The short answer is that I have tried. Since the 1990s I have recruited artists to work on various projects of mine, usually offering them about half of the income from the work while they also kept the rights to their art. The artists initially start off excited and promise they will have the art done on time. And they never finish. They always have good reasons: they are getting divorced, they are graduating, they started a full-time job, they have a commercial project to work on, they are going through things, and so on. I do think that maybe I am the problem, but when I ask people, they assure me that I am not. Which just might mean that I am a big problem. In contrast, AI image generators deliver the images I need reliably, which is why I use them. If I could find a reliable artist who can keep up with my writing schedule, I would be happy to give them 50% or more of the royalties, which takes me to the last reason I use AI image generators.
People who played Call of Cthulhu in the 1990s often remember me as the guy who gave away about 50 CoC adventures via the web. These days I write Mike’s Free Encounters, create Mike’s Free Maps and all my other new content is Pay What You Want. I remember saving my meager minimum wage income as a kid so I could buy games, so I am committed to making game content available for free when I can. I also did a failed Kickstarter promising 99 books for 99 cents each; so I have some relic 99 cent books out there as well. It seemed like a good idea at the time.
While we are unfortunate to live in a brutally extractive society ruled by oligarchs, I am fortunate to have a full-time job and can use this to help other people in various ways, including free gaming material. I would never expect an artist to donate art to my projects so I can keep giving them away and hence using an image generator makes sense to me. I would, of course, love to team up with an artist who would want to work on free projects, but it would be insane to expect that from anyone. And that is the explanation of why I use AI images in my D&D content. Am I wrong to do this? Obviously. As I said, like my fellow humans, I do a lot of bad things. I will write about why AI image generators are bad and probably also reveal other ways I am a bad person, like how I own a truck and use electricity.
A Philosopher’s Blog is Now on Substack!
You can subscribe and read for free.