I use AI artwork generated using the Bing Image Creator, StarryAI, and Stable Diffusion. I also use artwork from Pixabay and Pexels. You may even find some art from public domain repositories.

I prefer using graphics that I created. If what I need is beyond my skills, then I look for sources that are freely available for me to pull from. If I still cannot find what I am after, then I turn to AI as my last resort.

I recognize that there are concerns surrounding the use of AI imagery. For that reason, I want to be fully transparent in how I use the tools and all of my efforts to ensure that the art I use is not infringing on work that someone else owns.

There are a few reasons why I use generative AI for some of the art in my posts.

I’m No Artist While I can hold my own with an encounter map, I am not an artist. I cannot draw landscapes and buildings and objects. I most definitely cannot draw people. Since such images help to tell the stories I cover in my posts, I must turn to other sources for those images.

Copyright Rules Many will complain that some generative AI violates copyrights. I will get to some of that later on, but the bottom line is that if I were to use a piece of art from the Player’s Handbook or the Dungeon Master’s Guide as a thumbnail or section header in my posts, then I most definitely would be violating a copyright. I do not own those images and I cannot use them in content that sits behind a paywall. Sure, Wizards of the Coast might not care too much about my piddly newsletter today, but that doesn’t make it right. They own that art and I should not be using it. This is why I am actively moving away from using that art in my content except in cases where I can conform to the fan use policy.

Cost I am not a mainstream publication. I am just a guy that likes D&D and wants to make some content about 4th edition. As much as I would love to commission an artist to build me a library of assets that I can use, I simply do not have the resources to do this. If I get enough paid subscribers and PayPal contributions, then hopefully I get to a point where I can commission some real art. But, until then, I am just a guy with a hobby using the tools available to him.

Art is Not Featured My content is about playing 4th edition, running 4th edition, and the history of 4th edition. As previously stated, I am not an artist. I do not have posts showing off any of the artwork that I have used. It is not what I do.

Yes. AI art is obviously AI art. I am not fooling anyone nor am I trying to fool anyone. In fact, I am glad that the AI art is so obviously generated by AI. I don’t want anyone to think I am trying to pass off this art as something I created. I will actually sometimes even use the AI art that is comically AI generated. I am not trying to prove anything.

Keep in mind that the AI art I am using is mostly just to keep the posts from being a wall of text. It is an accent and nothing more. I might use it for a thumbnail or as a section header to help me capture the spirit of what I am writing about. But this art is never the focal point of what I am creating. It is meant to be glanced at as you make your way to the next section and that is it.

There are some who will say that using artwork from the official books is just fair use. Since I am only using it for a thumbnail or an accent, then that is fine. This is not true. In fact, fair use is quite the opposite.

Fair Use applies when I am sharing a piece of copyrighted material and offering some analysis or critique of it. For example, when I post a power card from the Player’s Handbook and then provide my thoughts on how to use this power and what tactical advantages it has, this is fair use. I am not simply republishing the power card. I am sharing it for the purposes of providing my analysis. If I shared that same power card as the thumbnail for a post without ever mentioning that power card in the post, that is most definitely NOT fair use.

The same is true for any image. If I am just using the image as the thumbnail or a section header, then the image is serving no purpose in my post except to add flavor. That is not fair use.

Here is an example from the music world. Weird Al Yankovic records a song with the same melody as Michael Jackson’s Beat It, but he uses his own, satirical, lyrics. This is fair use. He is creating a new work that is a parody of the original work. He does not need permission from Michael Jackson (though he still asked for it) nor does he owe any royalties.

But when MC Hammer uses the bass line from Rick James’ Super Freak in U Can’t Touch This, that is NOT fair use. The bass line was not being parodied. It was not being analyzed. It was simply being used in this other work.

---

Now that I have covered why I use AI generated art, I want to talk a little about HOW I generate it.

Generating the art for my content is no simple task. It is not a case of entering in a few words for a prompt and taking the first thing that comes back. I take great care to ensure that the art I use is unique and of the highest quality.

To that end, I follow these 4 steps with each and every piece of artwork I use.

I never reference specific artists or image sources in my prompts. I do not include individual artist names nor do I reference repositories such as Artstation in my prompts. I do not even use D&D or other product names in my prompts. My prompts focus strictly on the content I am looking for and the style of the art I want.

Since I first began researching AI art, I’ve done extensive investigation into different art styles and prompts that help me to properly describe the work I am looking for without needing to reference the work of other artists.

The art that I get is based purely on objective descriptions of the output I desire. I rarely provide any source images for the AI to start with. On the occasions when I do provide an image as a source for the AI, it is ALWAYS an image that I either own the rights to or one that was created by a prior iteration through this process I am describing here.

As much as is possible, I try to ensure that the art I am generating is original.

For every piece of art that makes it through this process, there are dozens, or even hundreds that did not. And for every piece of art that makes it into the finished product, there are a significant number that did not make the cut.

This is to say that I am by no means using AI art as a “shortcut”.

AI art is a tool to help me deliver the product I am after. I have high standards and if the tool is not providing the results I need, then that work is not used.

I spend hours (even days) on each and every image. Just as an artist may go through several drafts of their work before they are satisfied, I will continually refine the prompt that I am using until I get the images I am looking for. And the resulting images are then further filtered down until I have exactly the one I am after.

This is, perhaps, the most important step in this process.

I go to great lengths to ensure that any art I use is truly an original work. If I find that any art I create resembles an existing piece of art a little too closely, I disregard it and refine my prompts to avoid that situation again.

Anyone that has seen my videos and posts regarding the D&D licensing will know that I have no desire to “steal” content that is someone else’s creation. I take great pains to avoid approaching any gray areas with regard to copyright.

With every image that makes it this far, I perform an image search in Google and look through the results to ensure that the image I had was truly something unique. I am not just looking for an exact match. If I find something that I feel is too similar to what was generated, I will scrap my image and try again.

All art is derived from what came before. We all take inspiration from things we see and hear and read. AI art is no different from any artist that applies what they know to create something new. Using existing art as a basis for new work is not an infringement on copyright as long as the new work is truly a NEW work.

The images provided by the tools are rarely used without some modifications. I take the images through additional steps using tools such as GIMP and Adobe Illustrator to enhance, refine, and alter the image.

Sometimes I may change the colors. Other times I might crop out a small portion of the image. Other times I might edit the image to remove imperfections.

Every image I use will have some measure of my own style applied.

---

I share the concerns about copyright infringement and I hope my process described above helps to ease those concerns for the work that I am creating.

I also understand concerns about AI competing with aspiring artists. But we are all competing. I am competing with larger creators that have far more resources available to them than I have. The competition today is greater than it was in the past. That is just how such markets work.

New tools to simplify the creation process (or any endeavor) are constantly being developed. It is fair to ask creators what tools they used for their creations. I do not hide the fact that I use AI. It is a tool that I use very carefully.

As long as the end result is an original product, then I do not feel we should be concerned about the tools the creator used. At the end of the day, it is a question of ensuring that the product is unique and of high quality. That is my goal with everything I do.

Ultimately, it is up to you to decide if my use of AI to generate art is acceptable to you.