Starting this blog required a lot of planning, a lot of researching, a lot of learning. Just when I thought at least I had the storytelling part down, I realized I needed to polish most of my original writings while trying to present every entry mostly through narration, and then add the behind the scenes and dice rolls inserts, while topping most posts with a final commentary.
I had anticipated most of this when embarking in this adventure, but one thing I definitely did not foresee was the amount of art needed. Whether it’s a picture, a drawing, a map, an original artwork from Free League or something found on the web, every post needs at least one piece of art. That includes the about pages and the main header for the blog.
Ideally, I want every image to have a connection with the post it’s used in, but at times the subject of the narration is so specific, that it’s hard to find something fitting. This is where I had my first encounter with (relatively) new AI image generation tools: AI(-ish) programs that can render an image starting from a prompt given by the user, and with a bit of tweaking and patience the result can range from satisfying to good. Ideally these programs can generate an image tailored to my needs.
The problem is in how these programs operate: they sift the internet in search of specific pictures containing the details of the particular prompt the user has typed. Basically, the program looks for pictures corresponding to each and every word in the prompt, then puts everything together in a (mostly) cohesive way. This means that the program is almost certainly using details from existing artwork made by flesh and blood artists without any consideration for copyright, licensing, or royalties.
The foreseeable long-term consequence of this is that today’s real (human) artists are slowly but steadily training programs that could someday replace them. And I bet most of them have not agreed to this.
And so I made my decision: I won’t use any AI generated images for my posts and I will shows either images taken from real-life photos, real artists’ creations (with due credit given) or images of my own making (which I think will be mostly combat maps).
There will be a single exception to this rule: it’s far easier for me to visualize a character (PCs and NPCs alike) and recall it for a later moment if I have a picture/portrait of them. My artistic talents are lacking – insert laughter – in this department and for privacy reason I don’t want to use pictures of people taken from the internet. I will therefore use AI image generation tools to create believable portraits for my cast of characters.
There’s a degree of discomfort in this compromise, so I’ll keep the door open for a policy revision down the road.
Aaron C. Loop
