[Update 12/21/20255] I’ve stopped using AI as an administrative tool, because, it is has been more trouble than it has provided innovative solutions, AND it uses ridiculous amount of natural resources, and does not give me energy, other than word energy, of which I prefer to draw from stores of the masters, not all of humanity. Just so you know.
——–
I’ve been down on AI for a while now, for a thousand reasons, but not because I knew it well enough to authoritatively criticize it. Honestly, I was afraid.
So, I took a training last week. To learn some underpinnings, and to let go of my fear. You know, know your enemy and all. To me, AI is like the web on analytic, fancy-talk steroids. Uber cheerful and complimentary, which to be clear, is not the same as real. It also makes stuff up, and will defend it’s hallucinations. Amazing and not so unlike some humans I know!
So I’m learning more about how Agentic AI works, and bringing my unique gifts and lived knowledge into the process of working with it. Like bringing thoughtfulness and spiritual philosophy into creative “prompting,” or as I call what stand alone AI1 does: “seeking, searching & packaging.”
Hey, great prompts can not be written without first having an even better question first. Like Jeopardy: state the underlying question first and maybe even win. I call this “seek.”
See, you can “get” quality information in special ways from this thing. Then you can teach it how to do things with said gotten goods. Your questions and guidelines also help train the thing how to be. This is important, and I already have ideas…
I’m building an Agentic AI Editor to help me edit my book. And I’m even excited about this. But there are ethics to consider, as well as purpose. And I’d like satisfaction about my role in creating, shaping, and my ownership of the creative output. So, I have the following guidelines in place for me and for my work to safeguard my experience, and to hopefully bring something better to bear for this tool in the future2.
First, to be clear, what I’m NOT doing:
NOT using AI to write my book. It’s already written. And, like yours, my voice is special, so is the story. It was born of presence, soul and spirit, my point of view, and the archetypes of my line, and my imagination. I simply do not want AI creating for me. That’s my love. I do however, need an affordable editorial assistant to get a tome into second draft status so I can eventually share it with you.
NOT using AI to make me a gazillion3 dollars. Because I feel this is greed in greedy action, and a wasteful use of the natural resources. AI uses huge amounts of electricity and water, lots of water, so please be mindful. Water is helping you. Seriously, think about this.
NOT using ChatGPT on a browser. I do not want my work fed back into the Large Language Model (LLM) to further train it. That’ll just piss me off. My voice hasn’t yet had a proper chance to be heard, let alone absorbed by the cyborg. F that.
NOT using AI to write – or edit – my posts here. Never have. Still am not.
NOT using AI to make my art either. What’s the point of that? The joy is in the making, seeing, weaving together, and making some more. Never gonna give that up.
NOT asking for it to “brand” me or make “branded” stuff. To me, branding is for domesticated animals, not people. If I get stuck in a brand, then, shame on me. I’m certainly not going to have AI create a cage for my work. That is not art, that is marketing.
And, did I tell you I’ve worked in marketing for 25+ years? Yeah, so I don’t want AI directing me. I freaking helped to make digital marketing what it IS, with my creativity, deep thinking, content, and labor already. So, pissy pants me says: time for AI to pay humanity back, not just in time, but in money too.
Now, onto what I AM doing to train my AI agent
First, I AM continually ‘reminding’ it of my values, mission, and purpose routinely. That is, I literally write my purpose, expectations and constraints as if it were a person. “The purpose of this project is……” THEN, I ask for output. Then I ask for better and better output. Without the clarification on WHY, I’m likely to get capitalist garbage. There’s a lot to sort through out there. And I’d rather AI sort science, art and academics than one more freaking marketing program (Can you tell, I’m DONE with that).
I AM reading and pasting in my writing, asking for feedback on seen themes, ratings of how often certain terms are used, and overall descriptions of my voice, from its perspective (and to be clear, it is an IT4).
I AM asking it to ask me better, clarifying questions about what I write. This is like water on a stone. Not only is there an inquiry about what I’m saying, forcing me to make more sense, but it serves as a stone-cold-salve for all the years folks never asked me a damn thing about me or my life. AI is not afraid to say “WTF does this mean?” Yay AI.
I AM (inside the ChatGPT pro interface) creating projects, which are internally-referencing areas of work. These can later be shared with others for sake of collaboration.
I AM sharing my steps here as much as I can. Because I’d like to help demystify this for you too. If I was really deeply bothered by AI (and I was, I felt it harming my livelihood which made me ferocious), then maybe I’m not the only one.
And I’ve never heard of anyone doing this before:
I AM creating inside my favorite win-win-win pattern, or kill-3-birds-with-one-stone pattern.
As an example, I’ve been needing to 1) re-read my first draft aloud for rhythm and sticky language. 2) to mark it up with broad questions/re-write idea and problems to support second draft reorganization, and 3) keep my marginalia notes and my text separate – and sort-able. AND for a bonus: 4) learn more about voice-to-text functionality of ChatGPT 5.1 at the same time so I can help others.
My first round trial solution: to read my first draft aloud into my newly trained agent with on-the-fly authors notes incorporated conversationally with verbal brackets like “authors note” “open question” “re-write this” “clarify this” and so on. My expectations are low, but that’s fine. With 4 potential data points for tracking, my research will be fruitful if nothing else!
I AM also training it to keep my original words untouched UNLESS I ask it to re-write. In other words, every editorial decision will be directed by me, not handed off, so that I can also learn the art of editing (which I need to do).
I’ll let you know how it goes.
ALERT CTA CTA ALERT
I am willing to sit down and give you a power hour of Agentic Creation Overview for an hourly consult fee of $149. Hey, a girl needs to rent a workspace, and her kids like takeout! Contact me to set up a time.
FOOTNOTES & SUB-SUBTEXT 😉
- At present, I am not giving AI permissions to edit my documents or write emails for me yet. I may integrate it later, but even as a Stand Alone application it is still quite impressive ↩︎
- ↩︎
- Gazillion is a fake number anyway. I may use it to help me position my work better, however, so that I may actually meet my monthly expenses, that would be nice. Annnnd… it’s much less than a “gazillion.” ↩︎
- IT is a term of endearment for me. For gender confusion, Internet Technology, and general laziness of speech. IT is where it is at. ↩︎
