Since AI arrived on the scene a couple of years ago, I've been working closely with SUNO AI to create songs. This began with an original lyric I wanted to put to music. After awhile, I discovered ChatGPT, and both reworked my website and explored prompted graphics with AI assistance. Then, Google's Gemini came forward as a reliable resource for exploring graphic prompts with multiple, quick iterations before taking a final prompt to ChatGPT for generating. All three AIs have assisted in creating the foundation for many additional lyrics for hundreds of AI-generated songs, in collaboration with me and with my direction, with all resulting lyrics now being protected through my membership with Broadcast Music, Inc. (BMI).
Toward the end of last year, as SUNO reached a licensing deal with Warner Music Group (WMG), short animations were also introduced to the platform. I spent much of the Yule season creating enhanced graphics and animations for seasonal music on my YouTube channel. At the same time, I discovered that streaming royalties for AI-generated music are not a viable income source, so I have removed all but three music albums from international streaming; these albums are the debut album, Metros y Murciélagos, Love Songs vol. i, and Cover Queens, which is scheduled for release this summer.
A self-published companion book of AI-assisted lyrics translations into rhyming English was created for Metros y Murciélagos. This was a hand-reworked project, which required extensive coordination with ChatGPT as well as substantial edits, visits to online rhyming dictionaries and thesauri, and a great deal of wondering whether it took more time to teach ChatGPT what I wanted in a rhyme than it would have taken to do it myself, as I had in the pre-AI era of my debut book p.s. Fortuna.
The YouTube enhanced videos will be continuing this year, with eight mini-playlists for rituals of Goddesses and intentions, begun last year as global releases, and transformed into YouTube publications. This series will be scheduled through the end of summer. Following those releases will be eleven songs of tarot keys, with new tarot card graphics that I'll be working on with ChatGPT. All of the original versions of these tarot key songs are at my YouTube in their legacy SUNO lyric video format. Some of the audio may be reworked for the new videos, and some may stay the same.
This blog post is the first of eleven planned posts about the tarot keys, beginning with 0, “The Fool”, shown in its final, approved and edited version at the top of this post. Like many generated images, it took some interaction with AI to get the final product, along with a tiny bit of editing on my part, as well as tapping into a bit of what I've been studying in and about tarot with Builders of the Adytum (B.O.T.A.), which has a traditional correspondence school; now available online, with courses that were founded in teachings that began around the 1920s in the United States.
Our first Fool, based on a prompt that covered the basics of graphics for this card, resulted in what looked like a dazed hobo, perhaps high on life, with a daisy somehow stuck to a pole rather than being held in the opposite hand.
The B.O.T.A. Fool is not a hobo. S/he is a somewhat androgynous character wearing a white headband, dressed in a fancy jacket, colorful leggings and boots, with a designer bag swinging behind on a long staff or pole, a white rose in one hand, and a yappy little white dog happily watching at foot level for the next step. So, I asked for a more refined fool with a nice jacket and cap, with a happy expression yet open eyes.
Well, yes, this is exactly what I asked for; he's got a formal jacket and cap, yet it's hardly an interesting image, and the daisy still is not being grasped in the opposite hand, although it has moved to being precariously held upright between a thumb and a finger.
I needed a better inspiration, and it suddenly came to mind! We could create each card image to represent a decade, from the 1920s when the B.O.T.A. teachings began, with eras spanning to the 2020s of now. The Fool would represent the 1920s, and what better way than with a 1920s era flapper in a sequined, fringy dress, wearing a fancy headband, holding a handbag in one hand and a white rose in the other, with a friendly little white dog clearly looking forward to the path ahead. Here, we have something interesting; and something ChatGPT seemed to relish creating, interpreting the dress in a golden color. Our first approved iteration for a card image:
So, at this point, we have a concept for the keys, and our first approved image. All that needs to be done is cropping and applying the floral E Sylvia & AI watermark I try to remember to put on all my final AI-assisted graphics. Because this process required substantial human involvement, from the first iterations to the inspiration and final iteration, I'm told that the entire process and the final curated package are copyrightable. This may be true, and only time will tell how true it is.
What I do know, when working closely in collaboration with AI, is that it's like an instrument or tool rather than a pure iteration machine. While loose guidance seems to work better than micromanaged prompts, it takes more than a quick prompt to create a final product with that human touch. Even if the eventual output is entirely AI-generated, that iteration often follows an interactive process that looks quite a bit like a conversation between peers. As the learning process is accomplished, it's possible that future keys in this series require fewer interactions or iterations. But maybe not.
We'll have a look at each key, as it's designed, with a separate blog post. Our next key will be 1 Magician.
