Additional Lessons for the Staff Online AI Course
*Four new lessons, written to the same standard as Lessons 1–8, and designed for a phone screen: short blocks, one idea at a time, practice handled by the live chat rather than long reading. Sourced facts are marked and verified (July 2026). A source note closes this file.*
---
Lesson 9 — Writing Prompts with AI (and a Free Way to Get Better)
*Mixed sourcing. The prompt-writing method is general good practice; the One Million Prompters programme is [SOURCED] and verified.*
**On-screen text:**
You learned the basics of instructing AI in Lesson 3. This lesson goes one step further — a simple, repeatable way to build a good prompt, and something surprising: **AI can help you write prompts for AI.**
**A simple recipe for any prompt — the four parts:**
Think of a good prompt as four short pieces. You don't always need all four, but the more you include, the better the result:
1. **Role** — who you want the AI to be. *"You are a Grade 4 science teacher."*
2. **Task** — what you want it to do. *"Write a five-question quiz on the water cycle."*
3. **Details** — the specifics. *"Multiple choice, with an answer key, simple language."*
4. **Format** — how you want it laid out. *"As a numbered list I can copy into Word."*
Put together: *"You are a Grade 4 science teacher. Write a five-question multiple-choice quiz on the water cycle, with an answer key, in simple language, as a numbered list."* That one instruction will get you something usable on the first try.
**The trick most people don't know — ask AI to write the prompt for you:**
If you're not sure how to ask, tell the AI what you want in plain words and ask it to *write the prompt*. For example: *"I want to make a fun spelling activity for Grade 2. Write me a good prompt I could use to get that."* It will hand you a polished instruction you can then use — or improve. This is the fastest way for a beginner to get good.
**Live practice:** try the four-part recipe with the assistant now — build one prompt for a real task in your subject. Then try the trick: ask the assistant to write a prompt *for* you, and compare.
**A free way to get much better — recognised by Dubai:**
- **[SOURCED]** Dubai runs a free programme called **One Million Prompters**, launched under the directives of **H.H. Sheikh Hamdan bin Mohammed bin Rashid Al Maktoum**, Crown Prince of Dubai, through the Dubai Centre for Artificial Intelligence. Its goal is to train one million people in exactly this skill — writing effective prompts.
- **[SOURCED]** It is **free**, **online**, **self-paced**, available in **English and Arabic**, and it gives a **certificate**. It's a single course of four modules, starting from the basics.
- **[SOURCED]** It is **open to everyone — no restrictions on age, location, or background** — so it suits you as a teacher, and it's also something you could point your older students or your own family toward.
> **If you'd like to go further:** you can register and start for free here — **https://dub.ai/en/omp/**. It's optional and separate from this course, but it's a genuine, government-backed way to sharpen the one skill that matters most with AI. Taking it would also put you a step ahead in a field Dubai is investing in heavily.
**Check your understanding (not marked):**
- What are the four parts of a good prompt?
- If you're stuck on how to ask, what can you get the AI to do for you?
---
Lesson 10 — AI Will Not Replace You (and Why That's the Point)
*Sourced lesson. The reassurance is grounded in [SOURCED] UNESCO guidance, framed for the certification client teachers.*
**On-screen text:**
Let's name the worry directly, because many teachers feel it and few say it out loud: *"Will AI take my job?"* The honest, evidence-based answer is **no** — and understanding why will change how you use it.
**What the experts actually say:**
- **[SOURCED]** UNESCO — the United Nations education body — has published guidance arguing that **teachers must remain central to education** even as AI spreads. Its position is that AI can *support* teaching by handling certain tasks, but the real risk to guard against is not teachers being replaced — it is teachers being *de-professionalised*, handing away judgement they should keep.
- **[SOURCED]** UNESCO's framework for teachers is built on a **human-centred approach that protects human agency and judgement**, rather than surrendering it to a machine.
**Why a teacher cannot be replaced by a next-word predictor:**
Think back to Lesson 2 — AI predicts likely words. It has no relationship with a child. It doesn't notice the student who's quietly struggling, doesn't know that Ahmed needs encouragement today and Fatima needs a challenge, can't respond to a classroom's mood, and cannot care whether a child learns. Those are the heart of teaching, and they are exactly the things AI cannot do.
**So what does AI change?**
It removes some of the *load* — the drafting, the formatting, the first version of a worksheet — so you have more time for the part only you can do: teaching the child in front of you. A teacher who uses AI well isn't replaced; they're **freed up**. That is the point of this entire course.
**The message to hold onto:**
You are not competing with AI. You are the judgement, the care, and the human relationship that AI will never have. Your job is to make AI serve that — never the other way round.
**Check your understanding (not marked):**
- In your own words, why can't AI do a teacher's real job?
- What does AI actually take off your plate, and what does that free you to do?
---
Lesson 11 — Using AI With Your Students, Safely
*Sourced lesson. The pedagogy point is [SOURCED] to OECD 2026; the rules carry over from Lesson 5.*
**On-screen text:**
Once you're comfortable using AI yourself, the natural question is: *should my students use it, and how?* This lesson gives you the principle — always inside the age rules from Lesson 5.
**First, the boundary (from Lesson 5):**
- **[SOURCED]** Generative AI is **prohibited for students under 13, or below Year 7.** For those classes, this lesson is about *your* use to support them, not their direct use.
- For older students, any use must be **supervised, disclosed, and never in exams.**
**The one principle that makes AI help rather than harm learning:**
- **[SOURCED]** The OECD studied this in 2026 and found something every teacher should know: **AI helps students learn only when it's guided by clear teaching principles.** If a student simply *outsources* the task to AI, their finished work looks better — but they don't actually learn more. The performance goes up; the learning doesn't.
That single finding tells you how to use AI with students: **not to do the work for them, but to support the thinking.** Good uses look like:
- AI as a **practice partner** — quizzing a student, explaining a concept a different way, giving hints (not answers).
- AI as a **feedback tool** — a student drafts their own work, then uses AI to spot weaknesses to fix themselves.
- AI as a **starting point they must build on** — never a finished product they copy.
Bad uses are the mirror image: letting AI write the essay, solve the problem, or produce the answer the student then hands in. That's the outsourcing the OECD warns kills real learning — and it's also the academic-integrity breach from Lesson 5.
**The teacher's rule of thumb:** if the AI is doing the *thinking*, stop. If it's helping the *student* think, you're on the right track.
**Check your understanding (not marked):**
- What did the OECD find happens when students just outsource a task to AI?
- Give one good way and one bad way a student might use AI for an essay.
---
Lesson 12 — Spotting AI in Student Work (and Designing Around It)
*Practical lesson. General good practice framed for the certification client; academic-integrity rules carry over from Lesson 5. Interactive.*
**On-screen text:**
Every teacher now wonders: *"Did a student write this, or did AI?"* This lesson gives you both halves of the answer — how to notice, and the more powerful move: how to set work that AI can't easily do for them.
**How to notice AI-written work (signs, not proof):**
No detector is reliable, so treat these as prompts to look closer, never as certainty:
- **It doesn't sound like the student.** Vocabulary, fluency, or style suddenly far above their usual level.
- **It's fluent but empty** — polished sentences that say little, or drift off the exact question.
- **Wrong specifics** — invented references, or facts not from what you taught (remember Lesson 4: AI invents these).
- **No personal voice** — generic where you'd expect the student's own examples, opinions, or class references.
**The honest limit:** these are signals, not evidence. AI-detector tools are unreliable and can wrongly accuse an innocent student. Never treat a "signal" as a verdict — talk to the student.
**The stronger move — design work AI struggles with:**
Rather than police AI, set tasks it can't easily fake:
- Ask for **personal reflection** — "how did *you* find this," "an example from *our* class trip." AI has no personal experience.
- Ask them to **connect to something specific you did together** — a lesson, a discussion, a local event.
- Use **in-class work and spoken explanation** — "explain your answer to me" reveals real understanding instantly.
- Ask for **process, not just product** — drafts, rough notes, "show your working."
**Live practice:** tell the assistant a task you currently set your students, and ask it to help you **redesign it** so AI can't easily do it for them. Build one AI-resistant version of a real assignment.
**Check your understanding (not marked):**
- Why should you never treat an AI-detector result as proof?
- Name one way to design a task that AI can't easily complete for a student.
---
Source note — additions
**Sourced and verified (July 2026):**
- One Million Prompters: free, online, self-paced, four modules, certificate, Arabic & English; launched May 2024 under Sheikh Hamdan's directives via the Dubai Centre for AI (Dubai Future Foundation); registration at https://dub.ai/en/omp/; **open to all learners regardless of age, location, or background** (per the official dub.ai FAQ — this resolves the earlier student-eligibility unknown for this programme).
- UNESCO guidance that teachers must remain central and that the risk is de-professionalisation, not replacement; UNESCO's human-centred teacher framework protecting agency and judgement.
- OECD Digital Education Outlook 2026: generative AI supports learning only when guided by clear teaching principles; if students merely outsource tasks, performance rises without real learning gains.
**Carried over from existing lessons (not new claims):**
- The under-13 / below-Year-7 prohibition, no-AI-in-exams, and academic-integrity rules (Lesson 5, sourced to the UAE MoE 2026 manual).
**Note on the Global Prompt Engineering Championship:** its student-eligibility was not separately confirmed here and is not asserted in these lessons; only One Million Prompters is presented as open to all ages, on the strength of its own published FAQ.