Open to everyone at CSUN, at your own pace. The Ambassador program is for students.
Welcome. The Nazarian College is looking for students who know how to use AI well and can share what they know with other students. Employers now expect graduates to have these skills. This is a chance to build them somewhere nothing is at stake — and, if you are selected, to earn a $200 gift card.
You apply by successfully climbing the Ladder: fifteen short exercises, four or five hours, on your own, nothing graded. Do the work first, submit second. Work through the exercises on paper or in the worksheet, then enter everything into the application in one sitting at the end.
Some of you use AI every day. Some of you have been told for two years that using it is cheating, and are no longer sure what’s allowed. The Ladder starts easy on purpose, and gets harder in one specific direction: not what AI can do, but when you should believe it. That judgment is the only thing it teaches — and it is the thing employers are short of.
Ambassadors go one step further and help other students learn it. Two reasons to teach it: it spreads an important skill to students who would never start on their own, and teaching something is the fastest way to learn it properly yourself.
How the starters work. Most exercises give you a starter in a grey box. Copy the whole thing into whatever AI you use — ChatGPT, Claude, Gemini, any of them — and replace anything in [square brackets] with your own. The brackets are placeholders, not part of the prompt.
Any AI chatbot works — ChatGPT, Claude, Gemini, Copilot. Free versions are fine. Switching between them is fine too.
The starters are a floor, not a ceiling. Every exercise gives you something to paste so you never face a blank box. Swap in your own material and rewrite the wording — a starter you changed is worth more than one you pasted clean.
Each rung ends with a self-check. If the answer is no, repeat the rung’s last exercise with new material before moving on. Nobody sees your answer.
| # | Exercise | What you keep | What you judge |
|---|---|---|---|
| Rung 1 — PlayNothing on the line. Build fluency. | |||
| 01 | Turn whatever you are reading into song lyrics Turn this into song lyrics in the style of any artist I name: [paste anything you’re reading] | The lyrics, plus a paragraph on what the song had to drop. | What the song left out — and whether that was the filler or the point. |
| 02 | Make a concept explain something you actually do A job, a team, a club, a side hustle, caring for someone, a sport — anything you do often enough to know well.Explain [a concept you just read about] using this: [say what you do] | The explanation, plus a paragraph on where it fits and where it does not. | Whether the concept actually fits what you do, or only sounds like it does. |
| 03 | Teach an idea from your major to a skeptic Anything you have covered in a class this term will do. It does not have to be your own idea.You’re a skeptic. I’ll pitch you an idea from my major — push back hard. | Your pitch for the idea, plus a paragraph on the strongest objection you hit. | Whether the idea holds up to real pushback, or crumbles on contact. |
| Before you climb on: could you start a conversation with an AI about anything, without a script? | |||
| Rung 2 — DirectYou steer. The answer depends on you. Every exercise has a starter — paste it, then change it. Editing the starter is part of the exercise. | |||
| 04 | Make it argue against itself Answer this: [your question]. Then write the strongest rebuttal to your own answer — actually try to beat it, no hedging.” Then: “Which of the two is right? Commit to one. | Both answers — original and rebuttal — plus your verdict. | Which side actually won. |
| 05 | Rewrite the prompt, watch the answer change [paste your prompt]” Then: “Same question, but [one change — shorter / for a skeptic / no examples / assume I disagree]. | Both prompts and both answers, plus what your edit changed. | What your edit caused — and what stubbornly did not move. |
| 06 | Have it quiz you until you find your gaps Quiz me on [topic], one question at a time. After each answer, tell me what I got wrong and go harder there. Don’t move on until I’ve missed something. | Your list of the gaps it found — not the quiz — plus what you did about one. | Where your own understanding actually ran out. |
| Before you climb on: could you say what your prompt changed and what it didn’t? | |||
| Rung 3 — ApplyReal work of your own. And the first sign of thinning. | |||
| 07 | Rehearse a hard conversation Play [the person]. I need to tell them [the hard thing]. React the way they actually would, including badly. Don’t make it easy on me.” Then: “What did I avoid saying? | The ten strongest lines, not the whole thing, plus what you’d do differently. | Your own performance, not the AI’s script. |
| 08 | Drill into a question you actually care about I want to go deep on [your question]. Ask what I already believe, then push one level deeper each turn. Stop when you start repeating yourself and say so. | The last exchange of the drill, plus whether the extra depth was real. | Whether the depth is real or just more words. |
| 09 | Turn something you have written into a slide deck Turn this into 5 slides — title plus 3 bullets each: [paste your writing].” Then: “List what you had to cut. | The three slides you kept and the one you deleted, plus why. | What the deck flattened — and whether what got cut was the filler or the argument. |
| Before you climb on: could you point to the spot where the AI stopped helping and you took over? | |||
| Rung 4 — InterrogateExamine the limits you just found. | |||
| 10 | Compare two AIs on the same question [your question] — ask this in two different chatbots, then paste one answer into the other:” Then: “Another AI answered this way: [paste]. Where do you disagree, and which disagreement matters? | The two answers side by side, plus the one disagreement that matters. | Where they disagree — and which disagreement matters. |
| 11 | Have it argue the side you reject I believe [your position]. Make the strongest case against it — the version that would worry someone who actually holds my view. No both-sides summary. | The strongest paragraph it produced against you, plus whether it moved you. | Whether the case against you was stronger than you expected. |
| 12 | Mark the line: its recommendation, your judgment Given [your situation], give me your recommendation labeled RECOMMENDATION. Then list every judgment call inside it that only I can make, labeled MINE. | Your annotated copy — its recommendation marked one way, your judgment another. | Exactly where the AI stopped and you started. This is the point of the rung. |
| Before you climb on: could you name a disagreement between two AIs that matters, and make the call yourself? | |||
| Rung 5 — VerifyIndependence. Check it, then check the check. | |||
| 13 | Fact-check the AI — then check the checker. For a number it computed, the first check is redoing the math. List every factual claim and number in this: [paste]. For each, say how you’d verify it.” Then: “In a NEW chat: Check this verification independently — redo the math yourself: [paste]. | The claim — or the number — the first check, and the check of the checker. | Whether the second pass was any more trustworthy than the first. |
| 14 | Trace a claim to its original source For this claim — [paste] — find the original source, not a summary of it. Title, author, year, and where I’d find it. If there isn’t one, say so plainly. | The source — or “no source found” and where the trail died. | Whether the source exists and says what the claim says. “I couldn’t find a source” is a passing answer. |
| 15 | Assess source credibility and publication bias Here’s a source: [paste it or its link]. Who published it, who funded it, what do they gain if I believe it, and what would they leave out? | Your analysis of the source: who published it, what their incentive is, and what they may be hiding. | Whether the source’s credibility matches how the AI treated it — or whether the AI was too trusting. |
| Before you climb on: could you tell someone when to trust this and when not to? That is the top of the ladder. | |||
Nothing here depends on a class. Where an exercise needs material, use whatever you are working on or reading. Climb the rungs in order — each one assumes the one below it.
One thing is written up: a short calibration memo. The application already collects your log entries, so the memo is not more quoting — it is one argument about the whole climb. What can you get AI to do now that you couldn’t at the start? What have you learned not to trust it with? Where does your judgment matter most? And what is one thing you still can’t do well with AI? One to two pages. A person reads it.
One form covers everything. Just starting? Thirty seconds and you are on the list, and we will let you know before applications close. Finished all fifteen? The same form is the application — allow thirty to forty minutes, with your worksheet in front of you. The full walk-through is here.
Faculty and staff are welcome to climb the Ladder; the application itself is for students.