Tock's AI Explorer Book
Find out how smart machines learn, and how to stay in charge.
What is AI?
AI is short for artificial intelligence. It means computer programs that learn patterns from lots of examples, so they can do jobs that seem clever.
AI can suggest a song, turn your voice into text, sort photos or write a story. It can look like magic. It isn't. It is maths, patterns and a huge pile of examples.
AI is a powerful tool, and tools work best when you understand them. This book gives you eight AI explorer skills, so you know what AI is good at, where it goes wrong, and why a person should stay in charge.
How this book works
Every chapter has the same four parts.
- Read it. One big idea, with pictures.
- Try it. An activity you play right here on the screen.
- Away from the screen. A game for paper, pencils and friends.
- Quick check. Three questions. Get them all right to earn a star.
There are 9 stars to collect. Got a question wrong? No problem. Read the hint and have another go. That's how thinkers learn.
Choose a chapter
Chapter 1 · AI skill one
Spot the AI
Not every clever machine is AI. A calculator is fast, but it only follows rules. So how can you tell the difference?
An ordinary program follows steps that a person wrote. "If the button is pressed, ring the bell." It does the same thing every time. That's me!
An AI program is different. Nobody writes all its rules. Instead, people show it thousands or millions of examples, and it works out the patterns for itself. That is called machine learning.
You have probably used AI today without noticing. It suggests the next video, guesses the word you're typing, filters junk email and helps maps find a quick route.
Does each machine probably use AI, or does it just follow rules?
Away from the screen: the AI hunt
You need: paper, a pencil and a grown-up to ask
Walk around your home and list five machines or apps. For each one, decide: follows rules, or learns from examples? Ask a grown-up what they think. Some are tricky, and that's a good conversation.
Quick checkEarn your star
Chapter 2 · AI skill two
Teach with examples
How did you learn what a dog is? Nobody gave you a rule. You saw lots of dogs, and someone said "dog" each time.
AI learns in a similar way. To teach a computer to spot cats, people show it a huge number of pictures, each with a label: "cat" or "not a cat". The computer hunts for patterns that the cat pictures share.
The examples are called training data. Teaching the AI with them is called training. After training, the AI can make a guess about a picture it has never seen.
Tock has never seen a Blip or a Zorg. Blips are usually short with long antennae. Zorgs are usually tall with short antennae. Show Tock labelled examples, one at a time, and watch its guesses on six mystery creatures improve.
In maths: it's a graph!
Each creature is a point on a graph, with height along the bottom and antenna length up the side. Tock guesses by finding the nearest example it has seen. AI is built from maths like this.
Away from the screen: be the machine
You need: a partner and about twenty small objects
One person is the teacher and secretly picks a rule, such as "things that are red" or "things that roll". The teacher sorts objects into a Yes pile and a No pile, one at a time, without saying the rule. The other person is the machine and must guess where the next object goes. How many examples does the machine need?
Quick checkEarn your star
Chapter 3 · AI skill three
Make a smart guess
Here is the most important thing to know about AI. It doesn't know things the way you do. It predicts.
Finish this: "Once upon a ___". You said "time", didn't you? You've heard that pattern so often that the next word is easy to predict.
A chatbot works like that, on a giant scale. It has been trained on an enormous amount of writing. When you ask a question, it predicts a likely next word, then the next, then the next. The answer can sound wonderful. But it was built from patterns, not from understanding.
In maths: chance
90% means 90 out of 100. The AI is saying "very likely a cat", not "certainly a cat". About 1 time in 10, an answer like this will be wrong.
Be the AI. Choose the word that is most likely to come next.
Away from the screen: one word each
You need: three or more people
Sit in a circle and tell a story one word at a time. Each person adds the word they think fits best. Notice how the story makes sense, even though nobody planned it. That's a little like how a chatbot writes.
Quick checkEarn your star
Chapter 4 · AI skill four
Check the examples
An AI only knows what its examples showed it. If the examples are lopsided, the AI will be lopsided too.
Imagine teaching an AI about fruit using only pictures of red apples. Show it a green apple and it gets confused. It isn't being silly. It has never seen one.
This matters a lot when AI is used with people. If a voice assistant was trained mostly on adult voices, it may not understand children well. If the examples leave some people out, the AI works worse for those people. That is unfair, and it's called bias.
Look at the examples each AI learned from. Predict the problem.
Away from the screen: the lopsided survey
You need: paper, a pencil and your class or family
Ask only the people at your own table to name their favourite lunch. Write down the winner. Now ask a much bigger, more mixed group. Is the winner the same? If an AI had learned from only your table, what would it get wrong?
Quick checkEarn your star
Chapter 5 · AI skill five
Check the answer
AI chatbots can be wrong. Worse, they can be wrong while sounding completely sure.
Remember, a chatbot predicts likely words. Usually that gives a true sentence. Sometimes it gives a sentence that only sounds true. It can invent a fact, a date, a book or a quote that doesn't exist. Because it isn't checking, it doesn't notice.
So treat an AI answer like a tip from a stranger. It could be useful. Check it before you rely on it. Look in a book, ask a teacher, or see if trusted websites agree.
An AI wrote these sentences. All of them sound confident. Which ones have a mistake?
Away from the screen: two truths and a glitch
You need: paper, a pencil, a reference book and a partner
Pretend to be a chatbot. Write three confident-sounding sentences about an animal: two true and one that is wrong. Your partner must find the glitch, then check it in a book. Swap over.
Quick checkEarn your star
Chapter 6 · AI skill six
Ask a clear question
An AI can't read your mind. It only has the words you give it. Fuzzy question in, fuzzy answer out.
What you type or say to an AI is called a prompt. A good prompt is like good instructions for a robot: clear, complete and in the right order.
Compare "Tell me about frogs" with "Explain in three short sentences how a tadpole turns into a frog, for a Year 3 student." The second one tells the AI exactly what a good answer looks like.
Choose the prompt that will get the more useful answer.
Away from the screen: draw what I say
You need: a partner, paper and pencils
Draw a simple picture and keep it hidden. Describe it to your partner using only words, while they try to draw it. Compare the pictures. What instructions were missing? Try again with clearer ones.
Quick checkEarn your star
Chapter 7 · AI skill seven
Stay safe and fair
An AI chatbot can sound friendly. It isn't a person, and it isn't your friend. It is a program run by a company.
So keep your private information private. Don't type in your full name, your address, your school, your passwords or photos of yourself. If you wouldn't tell it to a stranger, don't tell it to an AI.
Being fair matters too. If an AI writes your story, you didn't learn to write a story. Using AI to get ideas or to check your work can help you learn. Using it to skip the thinking can't. Your teacher's rules come first.
Is each thing fine to type into an AI chatbot, or should you keep it private?
Away from the screen: our AI agreement
You need: a big sheet of paper, markers and your family or class
Make a poster with three headings: "We keep private", "We always check" and "We ask first". Write two or three promises under each. Everyone signs it. Put it near the computer.
Quick checkEarn your star
Chapter 8 · AI skill eight
You're in charge
AI is a tool, like a calculator or a hammer. A tool doesn't decide what to build. You do.
AI is quick at some things: finding patterns, sorting, summarising, suggesting. People are good at different things, and they are the things that matter most.
The best results come from teamwork. The AI suggests. You check, choose and decide. If something goes wrong, "the AI told me to" is never a good excuse. The person in charge is responsible.
Could AI help with this, or does it need a person to decide?
Away from the screen: design a helper
You need: paper and coloured pencils
Invent an AI helper for your school. Draw it and write three jobs it could do. Then write three things it must never decide by itself, and who should decide them instead.
Quick checkEarn your star
The last chapter · Put it all together
The big AI challenge
You now have all eight AI explorer skills. Here they are, with the question each one asks.
| Skill | Ask yourself |
|---|---|
| Spot the AI | Does it follow rules, or learn from examples? |
| Teach with examples | What examples did it learn from? |
| Make a smart guess | Is this a prediction? How sure is it? |
| Check the examples | Who or what was left out? |
| Check the answer | Have I checked this somewhere I trust? |
| Ask a clear question | Did I say exactly what I need? |
| Stay safe and fair | Is this private? Am I being honest? |
| You're in charge | Who decides, and who is responsible? |
Before you go, show what you know. Answer all ten questions to earn your last star and your certificate.
Final testTen questions, every skill
Certificate of AI Exploring
This certificate is proudly given to
for learning how AI works, where it goes wrong and how to use it wisely.
Look it up
Word bank
Every big word in this book, in plain language.
- AI
- Short for artificial intelligence. Computer programs that learn patterns from examples to do jobs that seem clever.
- Bias
- When something leans unfairly one way, often because the examples were lopsided.
- Chatbot
- An AI program you can talk to in words. It predicts likely words to build its answers.
- Data
- Information, such as numbers, words, pictures or sounds.
- Fact-check
- To check a statement against sources you can trust.
- Judgement
- Deciding what is right or best after thinking it through.
- Label
- A tag that says what an example is, such as "cat".
- Machine learning
- A way for computers to find patterns in examples instead of being given every rule.
- Pattern
- Something that repeats in a way you can predict.
- Predict
- To make a careful guess about what comes next, using patterns.
- Private
- Only for you and the people you trust.
- Prompt
- The question or instruction you give an AI.
- Training
- Teaching an AI by showing it lots of examples.
- Training data
- The examples an AI learns from.
Parents, carers and teachers
Notes for grown-ups
AI literacy means understanding roughly how AI systems work, where they fail, and how to use them safely and honestly. Children who understand that AI predicts from patterns are much better placed to question what it tells them.
No AI tool is needed for this book. Every activity runs on the page or on paper. Many AI products set a minimum age in their terms, so check before a child uses one, and stay nearby when they do. Chapter 7 is a good one to read together.
How to use the book
- One chapter per sitting. Each takes about 15 to 25 minutes, plus the off-screen activity.
- Ages 7 to 8 will get more from it with someone reading alongside. Ages 9 to 12 can usually work alone.
- Chapters build on each other, so going in order works best.
- Wrong answers are part of it. Every quiz question explains itself and can be tried again.
- Stars are saved in this browser on this device. Clearing browser data may reset them.
What each chapter covers
| Chapter | Idea | Where it helps at school |
|---|---|---|
| 1. Spot the AI | Rule-following programs and learning programs | Digital technologies, algorithms |
| 2. Teach with examples | Training data, labels, machine learning | Graphs and coordinates, classifying |
| 3. Make a smart guess | Prediction and confidence | Chance, percentages, patterns |
| 4. Check the examples | Bias from lopsided data | Fair samples in data and statistics |
| 5. Check the answer | Errors and made-up answers | Research skills, checking sources |
| 6. Ask a clear question | Writing clear prompts | Procedural writing, giving instructions |
| 7. Stay safe and fair | Privacy, honesty and age rules | Online safety, wellbeing |
| 8. You're in charge | Human judgement and responsibility | Ethics, decision making |