AI Policy and Usage

I am conflicted over AI use (my own and the general public's) and I am hoping that this working policy will help clarify my ideas.

Who this is for

This is specifically for my Digital Garden viewers/Digital Garden use.

Why I use it, and why I struggle with it

We can't and shouldn't do everything. I've learned this the hard way as I have been in the position of building multiple businesses from scratch and wearing all the hats. It's easy to say "pay people" but paying people to help sort through data, brainstorm ideas, and organize thoughts is not always feasible. In those cases, AI is useful as an assistive tool when and only when the proper checks and balances are in place. Is someone there to weigh real-world processes and uses against the generative logic that may miss or avoid nuance?

Ethical AI use begins with ethical societies and humans who understand collaboration as a community tool. Unfortunately, it seems the collective intent of this forced AI adoption is an increase profits by replacing human workforce with AI agents. The impact (environmental and economic damage which are hardest on the most marginalized) is not worth it. The end does not justify the means. We should be looking at how to use AI to enhance the quality of life not as how to skip/replace creative thinking.

How I use it

For me, the most important step in using AI has been working to devise and apply guardrails.

The start of every collaboration is my lived experience. My checks are in the process, not only after generation, although everything is reviewed and edited by me or my human executive assistant, but also in the briefing.

My process focuses on human-directed thinking through collaboration with the AI tool. I create briefs with explicit directions and instructions for engagement.

I do not ask it to think for me. Instead, I ask it to surface patterns in my thought process. To find discrepancies and to pinpoint weakness in my rationale. Not truth but rather sound arguments backed by valid, verifiable sources. Lead with disagreement. Challenge my thinking and force me to use valid arguments to back up those thoughts.

I credit sources for all ideas and verify any source suggested by AI.

Sometimes the collaboration is conversational brainstorming, with the directive that the agent push back and not placate. I mostly use this to refine my own thoughts and understanding while creating a working outline or plan.

AI suggestions are suggestions- nothing more. If I keep one, it is not a replacement for my thoughts but a rephrase that I consider and tailor until it represents me. It is still my writing.

Spelling: I have given permission for AI to correct obvious and blatant misspellings with a report of what was changed. All uncertainties are to be cleared with me. Quoted misspellings are never changed without my direction as those may be author's intent. See: Notes/Marginalia#^35a9d1

What AI does not do in my work

Credit

I credit Claude the way Stimpunks.World does. What Claude did, and where, is listed on my Marginalia page, and I update it as the work changes.

How Claude helped with this page

Claude, an AI model made by Anthropic, acknowledged the conflict that arose from my using it to draft this page. This statement serves to clarify how I used Claude specifically on this page. After I input a brief explaining my philosophy and rationale, Claude asked me questions in two rounds. Claude placed my answers under headings and it summarized the two Stimpunks pages I started from. The summaries provided a checkpoint. Claude also pointed out where my initial answers didn't match how I actually work, and fixed small typos. The words on this page are mine.

Keeping this current

This is a working policy. I update it monthly based on how I'm actually using AI.

I'll write about the public's use of AI, and its use in education, later.