I greatly the limit the use Generative “Artificial Intelligence” (aka, Large Language Models and their related technologies) on this blog. Effectively to the point that it will no longer be used for new content. I do not consider this to be a hard and fast rule. It is more a heavy rule that should cover 99.9% of future output.
I have used it.
I approach it mostly as a tool that is not for me, especially in the context of my own solo roleplaying.
tl;dr: I am not completely against the use of GenAI for people who are making their own solo roleplaying content (either to share or just personal use) but my own experiences with it have left me greatly preferring to not use it outside of a few experiments and occasions.
I wanted to write this page to describe my general thoughts about GenAI in solo roleplaying and especially this blog/project because the topic encroaches a lot into the solo-play space and people have different takes.
Some are 100% against using any AI product. Some are 100% for GenAI adoption to aid solo-play. Some use it to generate art while doing all the writing. Some use it to prepare tables and to enhance their writing.
Why I Stopped Using GenAI for The Doug Alone
It would be a lie to say that the pushback against GenAI was not at least some of the reason for why I started editing it out. It factored in. It just is not the only reason. Other reasons, more important ones, include:
- Opposition to GenAI’s overall practices (environmental, ethical, economical) to the point that I personally prefer to limit my use and would prefer my use to be less trivial than shortcutting the creation of art and oracles for my games.
- Challenging myself to find creative ways to express myself outside of my own comfort zone. Getting better at writing and finding ways to use human-supporting tools and resources to create the art in my projects.
- Joining in on a conversation between myself and other creatives where the books, images, music, et cetera used can be directly credited and link my art and output to the wide body of other works.
- Increasing my own creative control.
- Increasing my sense of ownership.
- Disliking a fair amount of GenAI-written RPG products and preferring to avoid using those.
- Finding that a lot of GenAI art looks bland and plastic and far too overfilled unless I spend a lot of time and effort to get it otherwise [a complaint that will likely be less true over time as the tools improve].
Overall, the main reason I stopped using GenAI except in rare cases is the fact that I find solo roleplaying to be a meditative act between myself and the various tools I use, finding a creative spark between my own interests and fairly niche products, and GenAI feels like an intrusion on that peace. That’s an overblown way of saying that my favorite aspect of solo play is using those oracles, tables, dice, cards, and books to create this odd and unique story based on algorithms that exist largely in my own mind and experience. Tools that shortcut that or offer to do that for me does not challenge me in the way I need to be challenged.
There is the potential wrinkle that some of the content, especially stock art, can be GenAI generated even if I try to avoid it. To demonstrate, I am going to use a source that actually does help you to find human created art. Pixabay (owned by Canva), a source for a lot of my earlier stock art use, has (as of 2025-10-09) 27K (labeled-as-)human-created pieces of artwork matching the phrase “Rainy Street at Night.” Which is frankly a phrase I might use a lot in my own games. They also have 13K (labeled-as-)ai-generated pieces matching the same search term. The wrinkle is that when you search for all art under that term, you get around 30K results. I’m sure there is a logic to it, but I appreciate the attempt. Note that some of the GenAI is fairly photo realistic. Sources that do not allow for that filter at all might result in easily using GenAI when the attempt is there to not. Unless you use older stuff, which I sometimes do precisely because of that.
Also note that image and content editors, including some industry standards, are now integrating GenAI heavily into their products to aid in editing and outright generating content. Adobe, Google, Canva, etc. Heck, WordPress (the backend of this blog) has increasing amount of AI tools even on personal installations like mine. I am simply not clicking that tab.
This is all to say that the line between “authentic” and “GenAI” content gets really fuzzy when you download an image that was a photo taken by a person but then edited by GenAI to touch up colors or tweak minor details and then uploaded as human-created. I am not going so far as to double- and triple- check all that.
How Do I Feel about Others GenAI Use in Solo Roleplaying?
Honestly, if it sparks joy for you, go for it.
Generating art for your personal game is all about you finding the place where you are happy with your game. Solo roleplaying can be quite difficult to get into the mindspace, especially initially. Generating tables or having an “GenAI GM” is about your comfort levels. If that’s how you like to play for whatever reason, it does not offend me. I used it at a one point to find out what I wanted and realized that I wanted was to not use it for myself.
I can imagine a lot of ways in which GenAI can meet people in the middle of the more “traditional” solo play does not work (e.g., rewording or explaining rules, tracking details, generating content to get over writer’s/players’ blocks). Technology tends to work best when it elevates people.
Note: this does not necessarily mean I want to purchase content generated this way, simply that I refuse to hate on people who are able to find their comfort zone through the help of GenAI.
If you post your actual plays and recaps using a lot of GenAI art, it will tend to detract from my enjoyment but I mostly read/watch those to see how people just enjoy themselves. If you are enjoying it then I am interested to see what you have made.
To give you an idea of how slight a detraction it would be: I would rather read/watch a session where the actual dice rolls (etc), gameplay, and content creation is included over one that has been turned into polished fiction. In this case, a person using some GenAI to illustrate an oracle result but including the actual oracle details would likely get my attention over someone who avoids GenAI but also avoids showings any of the gameplay.
I might buy a product that uses 100% human writing but is illustrated by GenAI. I have in the past.
I prefer to get something with no or few illustrations, though, over GenAI. Even goofy little doodles work for me.
I probably won’t buy a product which obviously (from samples) or openly uses GenAI to do the writing. I care more about the words than the pictures. Using GenAI to edit or to offer some suggestions is a bit different and I suspect a lot of people use it in such a way without necessarily outing themselves for it.
My most extreme take would be that I have zero interest in engaging in actual plays and recaps where the content was built by a GenAI gamemaster except in cases where it is posted as an open experiment or proof of concept. I am not hating on people who use it for their own enjoyment, just not my bag.
Where I Have Used GenAI for The Doug Alone
Two pieces of art somewhat frequent on The Doug Alone are GenAI, at least as a base. The first started out as a quick prompt of a man walking across a landscape with dice (which included a single die and domino because…GenAI). I wanted a quick avatar for an online forum. I ended up blurring, warping, and digitally painting over it until nothing of the original image technically remains but the core image was 100% generated by DALL-E 2:
The second image is slightly more hidden in this version and only shows up when I comment and in a few quick other places. It started as an attempt to try and get a GenAI to create an image of a haniwa. It did a very poor job. I kept poking it and adding more details and eventually this strange creation came out when I told it to make something like an “old haniwa in a Hawaiian shirt.” This was deep in my “only use GenAI ironically” phase.
It ended up looking so much unlike a haniwa and more like a terrible GenAI version of myself I adopted this as a anonymous image for online forums. I was still using it when I started the blog and have continued to use it because it is so odd. I both like it and hate it.
In the original versions of the early Eustace + Hitomi posts on the Prologue, I started out using GenAI to illustrate some key scenes. That’s also why I stopped using it. I just didn’t like it for my own creative process. It was too plastic. Too surreal. It added details that I did not need and avoided some that were crucial. I think the only remaining piece from that era is a cover for Ghosthearts and that actually was an in-joke that predated the blog. In other words, Ghosthearts existed in The Alabama Weird as a reference to an IRL conversation about cheesy 90’s young adult novels.
Other pieces of GenAI existed. Some still do, especially on older posts where I never got around to editing them out. To my recollection, the last piece of GenAI art was used in part 15 of Bleak + Pearl. I took an older GenAI image I had made of a crumbling tower and crafted it into that series’ standard two-to-three color stark-contrast style. I included a note on the end of the post that said “I am planning to generate no new AI content for this blog due to a lot of reasons.”
Which was a bit of a lie, since in a wholly different Eustace Delmont’s wholly different adventure, I did get ChatGPT to generate d66 table of Jordanian names. It was partially just an experiment of type to fill a gap in my many random tables.
After that, I thought it was interesting, but the idea that I had no idea what degree it was actually true enough to work meant I never really went back to it.