Technology · Unit 12

AI and Machine Learning

What artificial intelligence is and how it learns

Machine learning inverts normal programming: instead of writing the rules, you supply examples and the system derives rules that fit them.

The unit covers what AI is, machine learning, training data, neural networks, and the uses and limits of the technology.

This unit breaks down into 23 short steps and 121 questions, starting at difficulty 1 and building to 5. Below you can see exactly what it covers, how the path is structured, and worked examples with explanations.

Steps: 23
Questions: 121
Difficulty: 1-5

What this unit covers

Training Data
What AI Is
Machine Learning
Neural Networks
Uses and Limits of AI

Where this fits

Needs Programming and Data and Information.

Where people slip

A model reflects its training data, including its gaps and biases. A system that performs badly for a group usually saw too little of that group, and no amount of tuning fixes missing data.

How the unit is structured

AI and Machine Learning runs as 23 short steps that unlock in order. 17 are practice rounds and 6 are challenge rounds that pull together everything before them. Questions start at difficulty 1 and climb to 5 as you progress.

Step 1 · easierStep 23 · harder

Challenge rounds

Example questions

30 real questions from this unit, with the answer and the reason behind it, grouped by what they practise. There are 121 in the unit altogether.

Machine Learning

Fill the blankLevel 1
1. Machine learning lets a computer learn from ____ instead of being told every single rule.
- datacorrect
- paint
- sand
- music
Machine learning learns patterns from data.
True or falseLevel 1
2. In machine learning, a computer learns from data instead of being given every rule by hand.
Answer: True
That is the main idea of machine learning, learning from data rather than fixed rules.
Choose all that applyLevel 2
3. Which of these can help a computer learn in machine learning? Pick all that apply.
- Lots of example photoscorrect
- Past data about the taskcorrect
- Many labelled examplescorrect
- A cup of water
Example photos, past data and labelled examples all help a model learn, but water does not.
Multiple choiceLevel 2
4. What does machine learning help a computer do?
- Learn patterns from datacorrect
- Grow bigger in size
- Print faster
- Use less electricity all by itself
Machine learning lets a computer find patterns in data instead of being told every rule.
Sort into groupsLevel 2
5. Sort each thing by whether a computer learns it from data or follows a fixed rule.
Answer: Spotting spam email = Learned from data; Recognising faces = Learned from data; Adding two numbers = Fixed rule; Showing the current time = Fixed rule
Spotting spam and faces is learned from examples, while adding numbers and telling the time follow fixed rules.
Match the pairsLevel 4
6. Match each kind of machine learning to how it learns.
Answer: Supervised learning = Learns from labelled examples; Unsupervised learning = Finds patterns with no labels; Reinforcement learning = Learns from rewards and penalties
Supervised uses labels, unsupervised finds hidden groups, and reinforcement learns from rewards and penalties.

Neural Networks

Fill the blankLevel 1
7. A neural network is loosely inspired by the human ____.
- braincorrect
- knee
- ear
- hair
Neural networks are loosely based on the brain.
Choose all that applyLevel 2
8. Which words go with neural networks? Pick all that apply.
- Neuroncorrect
- Nodecorrect
- Layercorrect
- Sandwich
Neurons, nodes and layers are all parts of neural networks.
Match the pairsLevel 2
9. Match each neural network word to its meaning.
Answer: Neuron = A small unit that does a tiny calculation; Layer = A row of neurons; Network = Many neurons connected together
A neuron does a tiny calculation, a layer is a row of them, and a network connects many together.
Multiple choiceLevel 2
10. A neural network is loosely inspired by which part of the body?
- The braincorrect
- The foot
- The stomach
- The elbow
Neural networks are loosely inspired by how the brain's cells connect.
Odd one outLevel 2
11. Which word does NOT belong with neural networks?
- Garden hosecorrect
- Neuron
- Layer
- Node
Neurons, layers and nodes are parts of a neural network, but a garden hose is not.
True or falseLevel 2
12. Neural networks are loosely modelled on how brain cells connect to each other.
Answer: True
The design of neural networks was inspired by connected brain cells.

Training Data

Fill the blankLevel 1
13. To teach an AI to spot cats, you show it lots of ____ of cats.
- picturescorrect
- smells
- sounds
- weights
The AI learns from many pictures of cats.
Multiple choiceLevel 1
14. What is training data?
- The examples an AI learns fromcorrect
- The AI's power plug
- The screen it shows on
- The room it sits in
Training data is the set of examples an AI studies in order to learn.
Type the answerLevel 1
15. The examples an AI learns from are called training ____.
Answer: data
They are called training data.
Put in orderLevel 2
16. Put these steps of teaching an AI in the right order, first to last.
Answer: Collect lots of examples -> Train the AI on the examples -> Test how well it learned -> Use the AI to answer new questions
First you collect examples, then train, then test how it did, and finally use the AI.
True or falseLevel 2
17. If an AI is trained on wrong or messy examples, it can learn the wrong things.
Answer: True
Bad training data leads to a bad model, sometimes called garbage in, garbage out.
Odd one outLevel 4
18. Which of these does NOT help make training data better?
- Copying one photo a thousand timescorrect
- Collecting varied real examples
- Checking the data for bias
- Labelling examples carefully
Copying one photo many times adds no variety, while varied examples, bias checks and careful labels all help.

Uses and Limits of AI

Fill the blankLevel 1
19. When an app suggests songs you might like, that is a ____ made by AI.
- recommendationcorrect
- sandwich
- password
- battery
A suggestion like this is called a recommendation.
True or falseLevel 1
20. AI can make mistakes and is not always right.
Answer: True
AI is a helpful tool, but it can still be wrong.
Choose all that applyLevel 2
21. Which of these are real uses of AI? Pick all that apply.
- Chatbots that answer questionscorrect
- Recognising faces in photoscorrect
- Suggesting songs to playcorrect
- Digesting your lunch
Chatbots, face recognition and song suggestions all use AI, but your body digests lunch on its own.
Multiple choiceLevel 2
22. AI can help translate between languages, but what should we remember?
- It sometimes makes mistakescorrect
- It is always perfect
- It can never be wrong
- It fully understands feelings
AI translation is helpful but can still make errors, so it is good to double check.
Odd one outLevel 2
23. Which is NOT something AI can really do today?
- Feel true human lovecorrect
- Filter spam email
- Translate languages
- Recognise faces
AI can filter, translate and recognise faces, but it does not truly feel human love.
Put in orderLevel 4
24. Put these steps of using an AI answer responsibly in order, first to last.
Answer: Read the AI's answer -> Check it against a trusted source -> Correct any mistakes you find -> Use the checked result
You read the answer, check it against a trusted source, fix any mistakes, then use the checked result.

What AI Is

True or falseLevel 1
25. AI means computers doing tasks that usually need human thinking, like understanding speech.
Answer: True
That is exactly what AI is, machines carrying out tasks that normally need human intelligence.
Choose all that applyLevel 2
26. Which of these use AI? Pick all that apply.
- A chatbot that answers questionscorrect
- Face unlock on a phonecorrect
- A plain paper calendar
- A metal fork
A chatbot and face unlock both use AI to make smart decisions, while a calendar and a fork do not.
Match the pairsLevel 2
27. Match each short name to what it means.
Answer: AI = Artificial intelligence; ML = Machine learning; Bot = An automatic helper program
AI is artificial intelligence, ML is machine learning, and a bot is an automatic helper program.
Multiple choiceLevel 2
28. AI is mainly a part of which subject?
- Computer sciencecorrect
- Cooking
- Gardening
- Swimming
AI is a branch of computer science that makes machines act in smart ways.
Odd one outLevel 2
29. Which one is NOT really AI?
- A simple calculator adding 2 plus 2correct
- A chatbot
- A voice assistant
- A face unlock camera
A calculator just follows one fixed rule, while the others make smart choices using AI.
Guess the numberLevel 4
30. In which year was the term 'artificial intelligence' first coined, at the Dartmouth workshop?
Answer: 1956
The term artificial intelligence was coined around the 1956 Dartmouth workshop, seen as the start of the field.

Where these questions come from. Each unit starts as a plan of the concepts it should cover and the difficulty it should span. Questions are written against that plan with AI assistance, then checked by a validator that rejects anything without a single defensible answer, an explanation, or plausible wrong options. How we write questions sets out the whole process, and corrections are fixed in the bank and reach the site and the app the same day.

How you practise

This unit mixes 10 different question formats, so you are recalling and applying rather than recognising the same layout every time.

Choose all that apply
Fill the blank
Guess the number
Match the pairs
Multiple choice
Odd one out
Put in order
Sort into groups
True or false
Type the answer

Practise AI and Machine Learning

121 questions across 23 steps. Start with step one and crawl at your own pace.

Play this unit

Read about AI and Machine Learning

Explainers from our blog on what this unit covers. Each one ends with real questions from the bank.

What Is Machine Learning? A Beginner-Friendly ExplanationLearn how computers find patterns from data, how training differs from testing, and why model quality depends on examples.August 14, 2026 · 5 min read

More units in Technology

See all 18 units in Technology →