TECH CLUBBOZOMA INNOVATION HUB

Session 19 • Page 1

Learner Full Name:

School:

Assigned Workstation:

Class:

Session 19: Using AI Wisely & Safely

In Session 18 you learned that AI is a pattern-guessing machine that learns from examples. Today you meet the kind of AI everyone is talking about: chatbots that write answers to your questions. You will learn how to ask them good questions, how to catch their mistakes, and how to stay safe and honest when you use them.

1. Plain Talk: What Is an AI Chatbot?

An AI chatbot (for example ChatGPT, Gemini, Copilot or Meta AI in WhatsApp) is a program you can "chat" with by typing or speaking. It was trained on a huge amount of text from books and websites. From all that text it learned which words usually follow which other words.

When you ask a question, the chatbot predicts a good-sounding answer one word at a time, a bit like your phone keyboard suggesting the next word, but far more powerful.

Figure 19.0: A chatbot predicts likely words. It picks words that sound right, which is not always the same as true.

⚠️ AGE RULES AND SUPERVISION

Most AI chatbots say users must be at least 13 years old, and under-18s need a parent's or guardian's permission. In the BIH Tech Club, you only use a real AI tool with your mentor, at the Hub, or with a parent or guardian beside you.

Talking to AI: The C.L.E.A.R. Prompt

Session 19 • Page 2

1. What Is a Prompt?

A prompt is the instruction or question you type to an AI. A vague prompt gets a vague answer. A clear prompt gets a useful one. It is like sending a younger cousin to the market: "Buy food" brings back anything, but "Buy 6 eggs and a tin of tomatoes, under 50 cedis" brings back what you need.

2. The C.L.E.A.R. Prompt Formula

Letter	What it means	Example
C	CONTEXT: who you are and who it is for	"I am a JHS 1 pupil in Ghana…"
L	LIMITS: how long and in what shape	"…in 5 short bullet points…"
E	EXAMPLES & DETAILS: the facts it needs	"…with examples from farming and electricity…"
A	ASK: one clear task	"Explain how the Volta River helps Ghana."
R	REVIEW: read the answer and check it	Is it true? Is it what I asked? Ask a follow-up if not.

❌ Vague prompt

Tell me about rivers.

Which rivers? For whom? How long? The AI has to guess.

✅ C.L.E.A.R. prompt

I am a JHS 1 pupil in Ghana. Explain how the Volta River helps Ghana, with examples from farming and electricity, in 5 short bullet points.

3. Your Turn: Fix These Prompts

Rewrite each vague prompt as a C.L.E.A.R. one.

Vague: "Help with science."

Vague: "Write about my town."

AI Can Be Wrong: Check Before You Trust

Session 19 • Page 3

1. Confident but Wrong: "Hallucinations"

Because a chatbot predicts words that sound right, it sometimes makes things up: wrong dates, invented quotes, books that do not exist, even fake website links. People call this a hallucination. The answer often looks neat, polite and very sure of itself, so mistakes are easy to miss.

2. The 3 Checks

1

Compare with a trusted source: your textbook, your teacher or mentor, or an official website (for example one ending in .gov.gh or .edu.gh).

2

Does it make sense? If something sounds strange, too perfect or surprising, slow down and check it twice.

3

Find a second source: one website or one chatbot is not enough. Two trusted sources that agree are much safer.

3. Fact-Check Challenge

A pupil asked a chatbot about Ghana's history and got this answer. Underline the mistakes, then write the corrections. (Use your Social Studies knowledge or ask your mentor.)

🤖 "Great question! Ghana became independent from Britain on 6 March 1960. The independence leader was Kwame Nkrumah, who later became the country's first president. Today, Ghana's capital city is Kumasi, and its biggest lake is Lake Volta, one of the largest man-made lakes in the world."

Mistake 1 and the correct fact:

Mistake 2 and the correct fact:

Where could you check? Name one trusted source.

Deepfakes, Privacy & Honesty

Session 19 • Page 4

1. Deepfakes: When Seeing Is Not Believing

AI can now create photos, videos and voices that look and sound real. A fake one made with AI is called a deepfake. Scammers have used cloned voices to phone people: "Mum, it's me! I'm in trouble, send MoMo now!"

🚩 YOUR SESSION 15 S.T.O.P. SKILLS STILL WORK

Deepfake scams use the same tricks as phishing: panic, urgency and requests for money or PINs. Stop, then check another way: hang up and call the person back on the number you already know, or ask a family question only the real person could answer. Never share a shocking video or voice note until you know it is real.

2. Privacy: What You Type May Be Kept

Messages you type into an AI tool can be stored by the company and may even be used to train future AI. Treat a chatbot like a public noticeboard: never type your P.I.I. (full name with your school, home address, phone or MoMo number, passwords or PINs), and never upload photos of other people without their permission.

3. Honesty: Helper, Not Cheater

AI can be a patient tutor: it can explain a hard idea in a new way or quiz you before a test. But if AI does your homework and you hand it in as your own, you learn nothing, and it is cheating. Follow your school's rules, and if you used AI, say so.

How is the AI being used?	Helper ✅	Cheater ❌
Asking AI to explain fractions with a market example, then doing the homework yourself.
Copying an AI-written essay and putting your name on it.
Asking AI for 5 quiz questions to test yourself before an exam.
Using AI to answer test questions while the teacher is not looking.

Guided Hands-On Lab: AI Helper Check

Session 19 • Page 5

1. Lab Mission 1: The AI Helper Check (3D Lab)

Open the AI Helper Check 3D lab on your session page. It uses a pretend AI helper, so you don't need an account or the internet. Complete all three missions and record your results:

Mission	What happened?
Build a clear prompt. Which pieces made the answer better?
Check the AI's answers. How many of its answers were wrong?
Safe or never? Name two things you must never type into AI.

2. Lab Mission 2: Try a Real AI Tool (Only With a Mentor or Parent)

With your mentor, at the Hub, or with a parent or guardian beside you, use a real chatbot (for example ChatGPT, Gemini or Meta AI). No P.I.I., no passwords, no photos of people.

1

Write a C.L.E.A.R. prompt about something you are learning at school:

2

Summarise the AI's answer in one sentence:

3

Check one fact from the answer in a trusted source. Was it right?

4

Ask one follow-up question to improve the answer (for example "Make it simpler" or "Give me a Ghanaian example"):

📵 NO AI TOOL TODAY?

That's fine! Write your prompt in step 1 anyway. Your mentor will type some of the class's prompts into an AI tool on the projector, and you can check the answers together.

Session 19 Review & Verification

Session 19 • Page 6

1. Scenario Decisions: Wise or Unwise?

AI Scenario	Wise ✅	Unwise ❌
A chatbot gives you a date for a history project. You check it in your textbook before using it.
You type your full name, school and home address so the AI can "write a story about you".
A voice note that sounds like your uncle asks you to send his MoMo PIN quickly. You call him back on his usual number first.
You share a shocking video of a politician on WhatsApp straight away, without checking if it is real.

2. Written Review Questions

Question 1: Why can a chatbot give an answer that sounds sure but is wrong?

Question 2: What do the letters of C.L.E.A.R. stand for?

3. Self-Assessment & Mentor Sign-Off

Tick the column that reflects your confidence at the end of Level 9 (Artificial Intelligence):

Wise AI User Competency	With Help	Almost Alone	All By Myself!
1. I can explain how a chatbot creates its answers.
2. I can write a clear prompt using C.L.E.A.R.
3. I check AI answers against trusted sources before I use them.
4. I keep my P.I.I. out of AI tools, spot deepfake tricks and use AI honestly.

Mentor Signature: _____________________________

Date Completed: ____________________