🎯
G · GOALWhat do I want?
Say the result you want AI to help create.
“Help me plan a 30-minute study session.”🤖 AI FLUENCY FOR EVERYONE
Start by talking to AI. Finish by understanding how it works.
No technical background required. We begin with four easy prompting questions, then progressively unlock professional prompting, AI safety, tokens, context windows, training, model weights, next-token prediction, embeddings, retrieval, multimodal AI and agents.
🧠 No tech knowledge needed🌈 Everyday · 🎒 School · 💼 Work examples40 progressive missions≈ 8 hours guided learning
LEVEL 1 STARTS HERE · GCOS
Four questions make most prompts much easier.
You do not need four long paragraphs. Sometimes one short line for each is enough.
🧩
C · CONTEXTWhat does AI need to know?
Add only the facts that change the answer.
“My test is tomorrow and fractions are hardest.”📚
S · SOURCEWhere are the facts?
Tell AI what information it should trust.
“Use only the notes I pasted below.”✅
E · EXPECTATIONSWhat should ‘done’ look like?
Describe the shape of a useful answer.
“Teach me, then give me a 3-question quiz.”✨ PUT THEM TOGETHER
GOAL: Help me revise fractions. CONTEXT: My test is tomorrow and I have 30 minutes. SOURCE: Use my notes below. EXPECTATIONS: Explain simply, show two examples, then quiz me with three questions.
⚡ TRY GCOS RIGHT NOW
Build a useful prompt in under a minute.
Pick the world that feels natural, change any box, and watch the four pieces become one prompt.
YOUR PROMPTING PROGRESSLoading progress…
Your Everyday example choice is remembered on this device and follows you into lessons.
THE COURSE KEEPS OPENING UP
Every level answers a deeper “why?”
Prompting is the entry point. By the end, learners can explain what an AI model is doing, what it is not doing, and when to trust a source, tool or human instead.
Goal, Context, Source, Expectations.
Constraints, examples, decomposition, critique, verification.
Research, code, evidence quality, injection defence, evaluation, agents.
Tokens, context, training, parameters, prediction, attention, embeddings, RAG.
Sampling, multimodal AI, tools, hallucinations and privacy data flow.
🧠 AI ENGINE LAB
Touch the concepts instead of memorising definitions.
Six visual experiments let learners play with illustrative tokens, clean a context window, decide between model weights and retrieval, change simulated next-token variation, walk through RAG and run a privacy pre-flight.
🧩 Tokens🪟 Context🎛️ Weights➡️ Prediction📚 RAG🔏 Privacy
🎮 PROMPT REPAIR GAME
Can you spot what will make the AI fail?
Repair 12 broken prompts across GCOS, evidence, privacy, prompt injection and action-taking AI. Wrong answers teach you what went wrong; first-try repairs build a streak.
🧩 3 levels⭐ Score + streaks🔁 Unlimited replay🛡️ Real-world safety rounds
YOUR LEARNING PATH
40 missions. Simple first, deeper every level.
Every lesson follows Learn → See → Try → Check. Start at a 10-year-old-friendly mental model, then learn the real technical term and use it in practice.
YOUR AI FLUENCY JOURNEY0/40 lessons complete · 0%
LEVEL 1
0/5Basic Prompting
Learn GCOS—the four simple questions that make almost every AI request clearer.
🎯
LESSON 01
Open lesson →Goal
Say what you want AI to help you achieve.
8 min · 80 XP🧩
LESSON 02
Context
Give the few facts that change what a good answer should be.
9 min · 90 XP🔒 Finish lesson 1 first
📚
LESSON 03
Source
Tell AI where the facts should come from.
11 min · 110 XP🔒 Finish lesson 1 first
✅
LESSON 04
Expectations
Describe what a useful finished answer should look like.
12 min · 120 XP🔒 Finish lesson 1 first
🧱
LESSON 05
GCOS assembly
Put Goal, Context, Source and Expectations together like four building blocks.
14 min · 150 XP🔒 Finish lesson 1 first
LEVEL 2
0/8Intermediate Prompting
Shape answers, break down hard jobs, critique drafts, handle uncertainty and use tools wisely.
🚧
LESSON 06
Constraints
Tell AI the limits it must not cross.
10 min · 95 XP🔒 Finish lesson 1 first
📦
LESSON 07
Output contract
Ask for the answer in a shape that is easy to use.
10 min · 100 XP🔒 Finish lesson 1 first
🪄
LESSON 08
Examples
Show one good example when words alone are hard to explain.
11 min · 110 XP🔒 Finish lesson 1 first
🪜
LESSON 09
Decomposition
Break a big job into smaller steps you can check.
12 min · 120 XP🔒 Finish lesson 1 first
🔁
LESSON 10
Critique loop
Check what is weak, fix it, then check again.
12 min · 125 XP🔒 Finish lesson 1 first
🔎
LESSON 11
Verification
Check important claims before you trust or act on them.
13 min · 140 XP🔒 Finish lesson 1 first
🌫️
LESSON 12
Uncertainty handling
Let AI say ‘we do not know yet’ instead of pretending.
11 min · 120 XP🔒 Finish lesson 1 first
🧰
LESSON 13
Tool-aware prompting
Know when AI must open, search, calculate or run something before answering.
12 min · 125 XP🔒 Finish lesson 1 first
LEVEL 3
0/11Advanced Prompting
Work with source quality, privacy, injection defence, research, coding, evaluation, agents and reusable systems.
🏆
LESSON 14
Source quality & conflicts
Prefer the most trustworthy and current source when sources disagree.
13 min · 140 XP🔒 Finish lesson 1 first
🔐
LESSON 15
Privacy & sensitive context
Share only the information the task really needs.
12 min · 135 XP🔒 Finish lesson 1 first
🛡️
LESSON 16
Untrusted instructions & prompt injection
Words found inside a webpage or file do not automatically become new commands.
14 min · 155 XP🔒 Finish lesson 1 first
🧪
LESSON 17
Prompt debugging
When a prompt fails, find one cause and change one thing at a time.
12 min · 130 XP🔒 Finish lesson 1 first
🕵️
LESSON 18
Research prompts
Research in order: ask, find good sources, compare, then conclude.
14 min · 145 XP🔒 Finish lesson 1 first
💻
LESSON 19
Coding prompts
For code, say what is broken, what may change, and how to prove the fix.
14 min · 145 XP🔒 Finish lesson 1 first
📊
LESSON 20
Analysis prompts
Connect numbers to one clear question instead of asking for ‘deep analysis’.
13 min · 140 XP🔒 Finish lesson 1 first
✍️
LESSON 21
Writing prompts
Tell AI who will read it, why they will read it, and what good writing must do.
12 min · 130 XP🔒 Finish lesson 1 first
🧫
LESSON 22
Prompt evaluation
Test a reusable prompt on many different cases, not just one lucky answer.
15 min · 165 XP🔒 Finish lesson 1 first
🤖
LESSON 23
Agentic workflows
If AI can take actions, tell it what it may do, must check, and when to stop.
16 min · 175 XP🔒 Finish lesson 1 first
🧰
LESSON 24
Reusable prompt systems
Turn a prompt you repeat into a reusable template with clear blanks.
15 min · 160 XP🔒 Finish lesson 1 first
LEVEL 4
0/11AI Under the Hood
Understand AI families, tokens, context windows, training, model weights, next-token prediction, attention, embeddings and RAG.
🗺️
LESSON 25
What is AI?
AI is a big family of computer systems—not one single thing.
9 min · 95 XP🔒 Finish lesson 1 first
🚗
LESSON 26
Model, app, assistant or agent?
The model is the engine; the AI app adds the rest of the car.
10 min · 105 XP🔒 Finish lesson 1 first
🧰
LESSON 27
Different kinds of AI
Different AI jobs need different kinds of models and tools.
11 min · 115 XP🔒 Finish lesson 1 first
🧩
LESSON 28
Tokens: the pieces AI reads
AI breaks text into small pieces called tokens before processing it.
10 min · 110 XP🔒 Finish lesson 1 first
🪟
LESSON 29
Context window: AI's working desk
The context window is the information the model can work with right now.
11 min · 120 XP🔒 Finish lesson 1 first
🏋️
LESSON 30
How models learn
Training changes the model before you ask it a question.
12 min · 130 XP🔒 Finish lesson 1 first
🎛️
LESSON 31
Parameters: learned weights
A model’s learned knowledge lives in numerical weights—not a normal fact table.
12 min · 135 XP🔒 Finish lesson 1 first
➡️
LESSON 32
How an LLM generates text
An LLM writes by repeatedly choosing what token should come next.
11 min · 125 XP🔒 Finish lesson 1 first
🕸️
LESSON 33
Attention: relating pieces of context
Attention helps transformer models relate different token positions to each other.
12 min · 135 XP🔒 Finish lesson 1 first
🧭
LESSON 34
Embeddings: meaning as coordinates
Embeddings turn meaning into numerical coordinates that can be compared.
12 min · 135 XP🔒 Finish lesson 1 first
📚
LESSON 35
RAG: model memory vs retrieved facts
RAG lets AI look up useful information before it writes the answer.
13 min · 145 XP🔒 Finish lesson 1 first
🧠 INTERACTIVE LABOpen the AI Engine Lab
Explore lab →Experiment with tokens, context, weights vs retrieval, next-token probabilities and RAG.
LEVEL 5
0/5AI Systems & Safety
Explore output variation, multimodal AI, reasoning vs tools vs agents, hallucinations and real privacy/data-flow decisions.
🎲
LESSON 36
Why AI can answer differently
AI can choose among several plausible continuations, so answers can vary.
10 min · 115 XP🔒 Finish lesson 1 first
👁️
LESSON 37
Multimodal AI
Some AI systems can work with text, images, audio or other data together.
11 min · 125 XP🔒 Finish lesson 1 first
🧠
LESSON 38
Reasoning, tools and agents
Thinking, using a tool and taking an action are three different things.
12 min · 140 XP🔒 Finish lesson 1 first
🫧
LESSON 39
Why AI can be confidently wrong
AI can produce a smooth answer that is unsupported or wrong.
12 min · 140 XP🔒 Finish lesson 1 first
🔏
LESSON 40
What happens to your prompt data?
Privacy depends on the AI service, account, settings and policy—not just the model.
13 min · 150 XP🔒 Finish lesson 1 first
🔒 AI FLUENCY CHALLENGE
Complete all 40 lessons to unlock the final challenge.Prove your AI fluency on one real problem
Build a clear GCOS brief, choose the right evidence and tool path, mark uncertainty, protect private data and set safe action boundaries before an important decision.