Darrin Drader Designs, the publishing imprint of Darrin Drader, has used and will continue to use AI art. These services are an affordably priced alternative to traditional art, which we find to be of acceptable quality to illustrate our gaming supplements. To be clear, we believe in the capabilities of the technology, and we believe in the legitimacy of the technology.
As an individual, Darrin Drader has always been an early adopter of new technology. From his childhood obsession with early computers, to various hardware and software that have emerged throughout the years, he delights in experimenting, learning, and putting to use machines that have new capabilities. Thirty years ago, Photoshop was the “Big scary” new software that critics believed would put photographers and artists out of business. It didn’t. Twenty three years ago, the PDF became popular, and both traditional publishers and brick and mortar game stores were angry about its use, believing that it would put them out of business. It didn’t (in fact, Amazon did more damage to brick and mortar stores than the PDF ever did). These technologies were ultimately embraced by the very same people who originally spoke out against them and made new expressions of creativity possible that weren’t before.
We understand that there is some confusion about the nature of creative software and services that use machine learning, which we will attempt to address below.
On the accusation that AI art is stolen art: This is obviously the most serious allegation against AI art, but an understanding of how it works alleviates this concern entirely. AI art studies the full spectrum of human-created artwork so that it understands what things are supposed to look like. When you were a child, if you had never gone to the zoo and you heard about an elephant, you wouldn’t know what an elephant looked like except through children books, magazines, and television. Because your learning child mind was able to see what an elephant looked like, you were able to gain a comprehension of what they looked like, their size, how they move, and so on. AI is a neural network that is similar to the human brain in function. Unlike humans however, it is not capable of experiencing things directly. Instead, it relies on search engine data to inform it of what things are supposed to look like.
Further, human artists themselves draw inspiration from one another. In the fantasy field, Jeff Easley has stated that he drew his inspiration mainly from Boris Vallejo. Countless artists working today drew inspiration from him, Larry Elmore, Keith Parkinson, and Clyde Caldwell. As a result, much of today’s fantasy artwork possesses elements derived from those artists. Finally, the first thing that happens when anyone attends art school is that they study the works of other artists and learn how to replicate their techniques. This is the curriculum from which they learn. In other words, nearly all art in the world is derivative of the styles and techniques of those who came before them. AI learns in exactly the same way humans do.
On the accusation that AI art is “copied and remixed”: This goes back to the previous topic, but it’s worth addressing again. AI art is produced by a neural network. A computer neural network is defined as: “a method in artificial intelligence that teaches computers to process data in a way that is inspired by the human brain. It is a type of machine learning process, called deep learning, that uses interconnected nodes or neurons in a layered structure that resembles the human brain. It creates an adaptive system that computers use to learn from their mistakes and improve continuously. Thus, artificial neural networks attempt to solve complicated problems, like summarizing documents or recognizing faces, with greater accuracy.” In other words, this is not simply algorithm-based, and it is capable of far more than simply remixing elements. Instead, it seeks to understand the nature of of the problem or task and produce appropriate results
The authenticity of the neural network can clearly be seen by the fact that AI doesn’t always get it right. Whether we are talking about MidJourney or ChatGPT, it often produces content that is simply not a reflection of reality. The math is often wrong whereas standard computing always gets it right. Parts of images are distorted in interesting ways. These represent areas where the machine has not yet perfectly learned to create the things we expect. MidJourney, for a long time, created many-fingered hands, strangely shaped limbs, weapons that are held wrong or look wrong, and so on. If it were simply sampling and remixing, AI would have a much easier time getting these elements right because it would just copy/paste something from the original art that was done correctly.
On the accusation that AI will put artists out of business: This is a complicated issue that affects all areas of creativity, from art to writing to music, and it includes other disciplines as well. The sad truth is that many large businesses will use AI as a means to reduce their headcount and produce higher profits. On that note, publishers do not exist solely for the creation of jobs for artists. Artists provide a service that publishers need. If the publishers no longer need that service, that is an economic consideration.
That said, we believe there is an enormous difference between the use of AI art by a tiny publisher such as ourselves, who simply cannot afford to hire traditional artists, and large companies producing titles to be consumed by the mass-market who can absolutely afford the services of human artists. If large businesses don’t exist to provide opportunities for people then what is their purpose other than to harvest money from the economy?
We use AI art not because we are uninterested in helping human artists, but because we are not in a financial position where we are able to do so. All of the titles we currently sell that use AI art have the following characteristics: they are short products (15 or less pages), cost $1.99 (the same as we would charge if the product had no art at all), and have a shelf-life on the front page of DrivethruRPG that can be measured in hours, not days, weeks, or years. The average product at this time sells approximately 20 copies, which means that after DrivethruRPG gets their cut, we will make all of about $26 on the title. We spend more than two hours creating these titles, so our income is less than minimum wage, and a low price on a piece of original art would cost $250. Clearly that is not feasible and creates a barrier to entry into the market. We do this because we enjoy creating these products, not because they make a great deal of money
AI art allows small publishers such as ourselves to overcome the barriers to entry. While it is true that there is stock art available, that art often lacks the elements we need, and it is often not up to the style of quality we and our customers expect. To say that we should not publish if we can’t afford it is to say that people who are not involved with our business have determined that we cannot publish except under their terms. We reject these conditions.
On the belief AI will destroy the world: We hate to break it to people, but The Terminator was not a documentary. Please remove your tinfoil hat.
In conclusion: We have carefully examined the various issues from all sides and we have come to the conclusion that AI art is ethical. We also believe it is safe, and it will be a regular part of small publishing going forward, much as the PDF is today. In our products, we clearly indicate that they contain AI assets, we comply 100% with the rules regarding AI art on DrivethruRPG, and we do not attempt to claim copyright over that art (despite the fact that many of them have been substantially altered from the form that was originally generated). In fact, we are perfectly fine with it if people want to copy the art from our products and use it themselves, even without attribution, provided that they otherwise comply with the policies of the market place they are selling through.
Obviously we are not the final arbiter in any of these matters, and while we have faith that the court system will come to the same conclusions we have, we will of course comply with any final court decisions and restrictions that may or may not be handed down in the future.
Finally, we do support human artists to the best of our ability. When a product is large enough and has the budget for it, we do hire human artists. We ran the Reign of Discordia 2nd edition Kickstarter in 2022 and are currently employing artists to provide new illustrations for that product. We also currently support Dean Spencer’s stock art via Patreon with monthly financial donations, and we have no plans to eliminate that support.