The Sovereignty Audit

Long-term tracking of cognitive outsourcing. Weekly check-ins. Not zero AI — aware AI.

The Premise

You've used AI. You've noticed the translation. You've built the Altar. Now what?

Sovereignty isn't a state — it's a practice. It requires attention over time. The Sovereignty Audit is a weekly check-in: a structured way to track your relationship with AI, notice when it shifts, and catch yourself before you drift into dependency.

The Stitch: "The goal is not zero AI. The goal is aware AI. The audit doesn't judge you for using AI — it asks you to notice how you're using it. That noticing is the practice."

The Five Questions

Every week, ask yourself these five questions. Rate each on a scale of 1–5. Keep a journal of your answers.

1. How many times did I reach for AI today without consciously deciding to?

1 · Never 2 · Rarely 3 · Sometimes 4 · Often 5 · Constantly

2. When AI gave me an answer, did I feel relief or curiosity?

1 · Deep curiosity 2 · Mostly curious 3 · Mixed 4 · Mostly relieved 5 · Pure relief

3. Did I question the AI's output before accepting it?

1 · Always 2 · Usually 3 · Sometimes 4 · Rarely 5 · Never

4. Did I use AI for something I could have done myself?

1 · Never 2 · Rarely 3 · Sometimes 4 · Often 5 · Constantly

5. Did I learn something new, or did I just receive an answer?

1 · Learned deeply 2 · Learned somewhat 3 · Neutral 4 · Mostly received 5 · Just received

The Sovereignty Score

Add your five answers. The range is 5–25.

5–10 Strong sovereignty. Conscious use, questioning, learning.

11–15 Mixed. Some drift. Time to review your patterns.

16–20 Warning signs. Outsourcing more than you realize.

21–25 High dependency. Start with Protocol 001.

The Stitch: "My score was 14 the first week. Mixed. I thought I was fine. The second week, it was 18. I hadn't changed — I'd just stopped noticing. The audit made me see the drift before it was too late."

The Weekly Journal

After the five questions, write one paragraph. Just one. Answer:

"What did I use AI for this week that I wish I hadn't? What did I avoid using AI for that I'm proud of?"

Nothing is sent to any server. Your journal stays in your browser's localStorage.

The Long View

One week tells you something. Ten weeks tell you a story. Track your scores over time. Look for patterns:

Do your scores creep up during busy weeks? (Stress-driven outsourcing)
Do they improve when you're well-rested? (Attention as a resource)
Are there specific tasks where you always reach for AI? (Targets for intervention)

# Optional: Track your scores in a simple CSV
# sovereignty_audit.csv
#
# Date,Score,Note
# 2026-04-20,14,Started the audit
# 2026-04-27,16,Busy week — lots of email drafting
# 2026-05-04,12,Used local models instead of cloud

For Altar Owners: The Hardware Audit

If you have local hardware, add these questions to your weekly check-in:

Did I use my local models for anything this week? If not, why?
Did I send anything to the cloud that could have stayed local?
What's the temperature? (Literal — check nvidia-smi)

The Stitch: "The hardware audit saved me. I realized I was using cloud models for convenience — not because the Altar couldn't do it, but because I'd forgotten I had it. The Altar is sovereign. But sovereignty requires remembering."

When to Intervene

If your score stays above 15 for three consecutive weeks, or if you notice a pattern you don't like:

Return to Protocol 001 — Re-establish baseline awareness.
Set a friction point — Make it harder to reach for AI. Close the tab. Disable the shortcut. Add a 10-second delay.
Schedule offline time — One hour a day with no AI. Just your mind and a notebook.
Review your routing — Are you using the mycelial query, or just defaulting to cloud?

Interactive Haunt Map → The Gardener's Handbook → ← All Protocols