Train the Trainer

From learner to trainer

If you have completed an AI fluency course, you already have the foundation: the framework, the guardrails, and the judgment. This kit turns that foundation outward. AI is reshaping work and education right now, and the decisions being made today will shape both for years. The people closest to the impact have to be in the room when those decisions get made, and no algorithm can replicate the relational, human work at the heart of teaching and leading.

Deliver it in your voice. The kit gives you the structure; the credibility in the room is yours.

How to use the kit

Three pieces:

1. The session plan (part 3 below): the full workshop, section by section, with facilitation notes for every move: what to say, how long it takes, and what to watch for.
2. The exercises, embedded in the plan: warm-ups, the delegation matrix, and grade-band practice pairs at two depths.
3. The participant workbook (part 4): what your participants work through as you facilitate. Print it or share it; their answers stay theirs.

Preparation: read the session plan twice, run each exercise yourself once, and personalize the opening: one real, relatable AI moment of your own lowers the stakes for everyone.

The session plan

Shape: welcome and warm-up, Part A (the inner loop) with exercises, a 15-minute break, Part B (the outer loop) with exercises, then commit and close. Hands-on and judgment-first, not a tech lecture.

Welcome (20 min)

Outcomes to name out loud: participants will be able to define AI fluency, apply the 4D framework as a mental model, prompt and check outputs effectively, and connect AI fluency to professional skills they already have. Frame these as a promise for the session, not a test.

Learning norms, under a minute: use your voice, honor collaboration, participate actively, be flexible, respect differing opinions, share your gift. Invite the room to add one.

Three guiding rules, stated plainly: (1) no prior knowledge assumed, whether you have never opened a chatbot or use AI daily, you belong here; (2) use your organization's approved tool, because the skills transfer and what you are building is judgment, not just technique; (3) nothing that identifies a student, client, or beneficiary goes into any AI tool. Say rule three plainly, then pause. It is the non-negotiable.

Why AI fluency (10 min)

The literacy ladder: digital literacy asks "can I use this technology?"; AI literacy asks "do I understand what this is?"; AI fluency asks "can I work with this well, on my own terms?" Use the food analogy: you can know a lot about food without being a good cook.

Four myths to bust: digital literacy is not AI literacy; AI literacy is not AI fluency; you do not need to be a tech expert (fluency is built on professional judgment, critical thinking, and domain knowledge); and more use does not equal more fluency, because knowing when not to use AI is a core marker.

The framework (10 min)

The 4Ds in one line each: Delegation, what tasks are right for AI? Description, what do you need AI to do? Discernment, is the output high quality? Diligence, how do you take responsibility? Inner loop = Description and Discernment, the everyday cycle. Outer loop = Delegation and Diligence, the whether and how much.

Part A: the inner loop (45 min)

Prompting model, "a great prompt is like a good handover note": Role (who the AI should be), Task (the specific goal), Context (background and examples), Format (the exact output shape), Constraint (limits and guardrails).

Discernment, the CHECK method: Correct? Verify facts with outside sources. Human? Warm and authentic. Equitable? Free of bias and harmful assumptions. Curriculum-aligned (or standards-aligned for your field)? Kid-safe (or audience-safe)? Have the room say CHECK back once; spoken mnemonics stick.

Why discernment matters: hallucination is a plausible pattern not grounded in reality; sycophancy is the model telling you what you want to hear; bias is AI reproducing inequalities in its data. Facilitation note: research shows discernment behaviors are used least frequently when people engage with AI. Slow down here; this is the heart of the inner loop. Reassure the room: description is like planning a lesson or writing a handover, and discernment is like grading or reviewing. They already do this thinking.

Break (15 min). State the return time and hold to it.

Part B: the outer loop (75 min)

Two-minute re-entry: turn to someone new and share one thing from Part A.

Some work is the work because you do it. A reflective letter, a hard conversation, the thinking you do to understand why someone is struggling. AI taking it from you does not save time; it removes the point.

Diligence in practice, three moves: Verify (facts, reading level, anything that could mislead). Own (revise until it sounds like you and reflects your judgment, not the AI's defaults). Disclose (if AI helped produce it, say so; transparency is part of the professional standard).

Data protection block: use only approved tools; never paste anything that identifies a person (de-identify: "Student 1," "Client A"); assume inputs may be retained, and read the policies. Then the one-minute quiz, high energy: 60 seconds, list every kind of personally identifiable information you can. Rapid-fire review: names, ID numbers, dates of birth, addresses, photos, case status, health records, grades, contact details. None of it belongs in an AI tool.

Commit and close (15 min)

Real fluency means moving between the loops without thinking about which one you are in. That is not a failure of the framework; that is the framework working.

The commitment ritual: everyone writes one real task, not a hypothetical, that they will bring AI into this week, on a card. Collect them at the door or have people keep them; a written commitment is far more likely to happen. Close warm and specific, restate the commitments, and end on time.

Exit pulse against the four outcomes: comfortable teaching it, learned something new, or still unsure. Normalize "still unsure"; it names where to keep practicing.

The participant workbook

Give this structure to participants, on paper or as a document. It turns the workshop into something they do.

1. Before you start, three guardrails: no prior knowledge needed; use your organization's approved tool; never enter anything that identifies a person.
2. Warm up: the "Have you ever...?" checklist, plus one line: "One AI moment I want to get better at: ___."
3. The 4Ds at a glance: the two loops, four one-line definitions.
4. Activity 1, describe it then CHECK it: a real low-stakes task; build the prompt with Role, Task, Context, Format, Constraint; run CHECK on the output; note what you would change before using it.
5. Activity 2, the delegation matrix: sort the six seed tasks into the four quadrants, then add two of your own.
6. Activity 3, run the gates: a real task; what AI touches and what stays yours (before); how you will verify and disclose (after).
7. Commit and close: "This week I will ___" (real, not hypothetical), and the exit self-assessment: comfortable teaching it / learned something new / still unsure.