AI Fluency Fundamentals
The six 201 skills that separate sustained AI adopters from the 80% who quit. No coding. No tool-building. Applied judgment.
Bring with youDone before you walk in
- Laptop with one AI tool open. GenAI.mil if you have CAC; otherwise CamoGPT, ChatGPT, or Gemini. Confirm you can sign in before the session.
- A notebook or notepad. You will sketch a workflow map and seed your frontier sticky on paper.
- One real recurring task from your job in mind for Module 5 — something you do most weeks.
- Cameras on, mic ready. Two cold-calls, three show-of-hands, one silent timed exercise.
- If you’re a builder or supervisor, skim the Course 2 and Course 5 pages so you self-select the right next track.
Key termsThe vocabulary you’ll hear today
- The six 201 skillsQuality judgment, decomposition, context-building, iterative refinement, mode-switching, frontier mapping.
- CentaurYou design and verify; AI executes inside fixed phase boundaries.
- CyborgContinuous back-and-forth with AI; the boundary is fluid (drafting, prototyping, exploring).
- The frontierThe line between what AI handles well and what it doesn’t for your work — and which moves with each model release.
- The Red Pen ReviewMark every AI claim you would not sign — facts, references, internal consistency, procedure, substance vs style.
- The delegation equationIf you’d delegate it to a brand-new Marine, it’s an AI-shaped task.
Exercises in classWhat you will do live — and what “done” looks like
M2 · Quality Judgment (5 min). Two AI-drafted NAM citations for the same Staff Sergeant. Hands up: which one survives the awards board, A or B? Be ready to defend it before the reveal. Done: you can name why B (specific, verifiable facts) wins over A (filler adjectives).
M3 · The Delegation Equation (3 min). Pick one task you owned this week. On your notepad: who is the audience, what does done look like, what could go wrong? Decide — would you hand it to a brand-new Marine? Done: you have one yes/no with a reason.
M4 · Red Pen Review (15 min + 10 debrief). Three AI-generated documents drop into Teams chat. Ten silent minutes. Mark every claim you would not sign — references, facts, internal consistency, procedure, substance vs style. Done: 3–4 issues per doc; you can categorise each as fabricated facts, missing context, or tone mismatch.
M5 · Workflow mapping (8 min). Pick one recurring task. Sketch 3–5 subtasks. Mark each: only-me, centaur, cyborg, only-AI. Decide centaur or cyborg overall. Estimate time saved if AI handled the right subtasks. Done: one circled subtask you’ll change next week.
M6 · Frontier-map seed (5 min). Three columns on a sticky: Handles well · Handles poorly · Moving frontier. Two real examples per column. Star the failures that surprised you. Done: you can name one failure case worth sharing with your section.
Anchor phraseYou are not learning to use a tool. You are learning to manage one.What you’ll be able to doBy the end of the session
- Tell the difference between centaur and cyborg work, and pick the right mode for one task on your desk.
- Run the Red Pen Review protocol on any AI output before your name goes on it.
- Spot when a task is AI-shaped using the delegation equation.
- Sketch a workflow map for a recurring task and call out the AI-appropriate subtasks.
- Start a frontier sticky — what AI handles well, what it doesn’t, what’s moving.
HomeworkThree things. One week. Issued at Slide 51 — write it down.
- Use it once for real this week. Pick one task from your M5 workflow map. Use AI on it — centaur or cyborg, your call. Bring the result, success or failure, to your section meeting.
- Add to your frontier sticky. Two more entries: one task AI handled well, one it didn’t. We collect these for the unit map in Week 4.
- Builders & supervisors: read the Course 2 (Builder Orientation) and Course 5 (Supervisor) course pages so you walk into next week on the right track.
- If something blocks you (no tool access, unclear policy), email it to the program lead by EOW — we’ll resolve it before Week 2.