Day of AI Australia

2026 program

Lesson 1 · How Artificial Intelligence Systems Work · Years 3–6

A reminder to teachers

To get the most out of these lessons, please “ENTER FULL SCREEN” — press F.

Day of AI Australia
2026 program

Day of AI Australia acknowledges First Nations peoples as the Traditional Custodians and first scientists, innovators, and educators of this land and their continuing connection to Country.

We pay our respects to the Elders past and present.

Students in classrooms all across Australia are taking part in Day of AI, just like you!

🧑‍💻👩‍💻🧒

1,000s of classrooms nationwide

In Australia, and throughout the world, AI technology is all around us.

Throughout the series of lessons, we’ll be looking at some amazing case studies of how this exciting technology is already used in our homes, local communities and in our natural environment.

Years 3 – 6

Lesson 1:

How Artificial Intelligence Systems Work

2026 program

? ? ? ? ? ?

Have you heard of artificial intelligence before?
Have you used artificial intelligence?

What is AI?

🧒🥽

▶ Watch: an intro to AI

Artificial Intelligence (AI)

A program made by people that makes computers do things that seem intelligent (or smart) in the same way that humans are intelligent.

The 5 Big Ideas of AI

1It can understand its environment
2It can plan and make decisions
3It can learn new knowledge and skills
4It can interact with humans and the environment
5It has an impact on society

🧑‍🤝‍🧑Student led activity

AI or Not?

Let’s play a quick game to test if you can recognise AI!

▶ Play AI or Not AI?

ai-or-not.dayofaiaustralia.com

AI or Not AI?

Some things use AI and some don’t. Can you tell which is which?

Play!

History of Artificial Intelligence

The Idea of AI is Born

1950s

🌍

Early Optimism

1960s–70s

🤩

AI Winters

1980s–90s

📉

Big Data and Machine Learning

2000s–2010s

📈

Generative and Everyday AI

2020s

💰

How does a machine become intelligent?

Parts of Machine Learning

Let’s look at the basic building blocks of all AI:

DATA

ALGORITHM

PREDICTION

DATA

ALGORITHM

PREDICTION

Data

Data is information that we collect and record to help us understand something.

Data can be:

Numbers (time, views, temperature)
Text (novels, articles, blogs)
Audio (speech, music, animal sounds)
Images (photos, drawings)
Video recordings (movies, TV, home videos)

☝️ EXTENSION QUESTIONS:

Do you think you own data about you?

Do you think companies should be allowed to use your data, if their product or service is free?

DATA

ALGORITHM

PREDICTION

Datasets

A dataset is a collection of curated data.

Data can be:

Numbers (time, views, temperature)
Text (novels, articles, blogs)
Audio (speech, music, animal sounds)
Images (photos, drawings)
Video recordings (movies, TV, home videos)

In order for machines to learn, they need data.

A curated data set is one that has selected, organised and (possibly) labelled data.

DATA

ALGORITHM

PREDICTION

Algorithms

An algorithm is a procedure used for solving a problem.
Algorithms act as an exact list of instructions that do things step-by-step, like learn what a cat or dog looks like.
There are lots of unique algorithms in the world of AI. They are different for each and every AI.

DATA

ALGORITHM

PREDICTION

Predictions

A prediction is like a guess, but with more information.

Predictions are made based on prior knowledge.

This is part of learning, otherwise known as intelligence.

🤖✏️

f(x) = ℓ·m … x²−4x+5 ≤ 5

🧑‍🤝‍🧑Student led activity

Google Quick, Draw! The Data

Let’s test how a machine uses prediction.

✏️ Quick, Draw!

quickdraw.withgoogle.com

✏️🎨🖐️

🧑‍🤝‍🧑Student led activity

Google Quick, Draw!

Let’s test how a machine uses prediction.

✏️ Quick, Draw! The Data

quickdraw.withgoogle.com/data

📝Case studies

Giant kelp forests are vital marine habitats, but they’re disappearing due to increasing temperatures in oceans.

Google and partners are using AI, satellite data, and genetics to restore these forests.

Watch the video to see how technology is helping nature recover — and why this matters for ecosystems and communities.

Learning is the 3rd ‘Big Idea of AI.’

Like humans, AI needs to be ‘taught’ how to do things, before it can do anything useful.

Have you ever thought about how you learn?

5 Big Ideas wheel — Learning highlighted

🧑‍🏫Teacher led activity

How do we learn information?

This is a ‘Fleep’.

🧑‍🏫Teacher led activity

How do we learn information?

This is also a ‘Fleep’.

🧑‍🏫Teacher led activity

How do we learn information?

This is also a ‘Fleep’. What do all the ‘Fleeps’ have in common?

🧑‍🏫Teacher led activity

How do we learn information?

But — this is a ‘Bloop’.

🧑‍🏫Teacher led activity

How do we learn information?

This is also a ‘Bloop’.

🧑‍🏫Teacher led activity

How do we learn information?

What do all these ‘Bloops’ have in common?

🧑‍🏫Teacher led activity

How do we learn information?

Top row: three Fleeps. Bottom row: three Bloops.

🧑‍🏫Teacher led activity

How do we learn information?

Let’s test if you think this is a ‘Fleep’ or a ‘Bloop’?

Time to stand up!

Go to the left side of the room if you think it is a ‘Fleep’
Go to the right side of the room if you think it is a ‘Bloop’
How confident are you in your choice?

🧑‍🏫Teacher led activity

How do we learn information?

ANSWER: It’s a ‘Fleep’!

Why?

All the ‘Fleep’s have big eyes and fur.
Both had round or pointy ears.
All the ‘Bloop’s have small eyes and no fur.
This one has big eyes and fur — so it must be a ‘Fleep’!

Note you saw only 6 samples! AI can look at millions and billions of examples to learn!

AI training, testing and prediction doesn’t just apply with video, images or motion!

It can also be applied to sounds. 🔊

🧑‍🏫Teacher led activity

How do we learn information?

Play these two different sounds of native Australian birds.

🔊 This is the sound of a kookaburra

🔊 This is the sound of a cockatoo

🧑‍🏫Teacher led activity

How do we learn information?

Now that you’ve listened to sound examples of a kookaburra and a cockatoo, play the first 10 seconds of this new sound.

Do you think it is a cockatoo or a kookaburra?

🔊

Time to stand up!

Go left if you think it is a kookaburra
Go right if you think it is a cockatoo
How confident are you in your choice?

🧑‍🏫Teacher led activity

How do we learn information?

How do you describe what a cockatoo sounds like?
How do you describe what a kookaburra sounds like?
On a scale of 1–10, how confident are you? (1 is not confident and 10 is certain)
How did you reach this confidence score?
What extra information might you need to improve your confidence?

How important is the confidence rating? Would you trust an AI system without one?

🧑‍🤝‍🧑Teacher-led activity It's a kookaburra

It’s a kookaburra! Note that you heard only 2 samples! AI can listen to millions and billions of examples to learn!

🧑‍🤝‍🧑Teacher-led activity

What do you think would happen if this AI heard a rooster?

Remember, the AI has only been trained to recognise kookaburras and cockatoos.

It’s now time for you to train your own machine to become intelligent!

Recap of Image Recognition

INPUT

⭕

An image

➜

LEARNING ALGORITHM

OUTPUT

Recap of Image Recognition

INPUT

⭕

An image

➜

LEARNING ALGORITHM

Learn features from training data

(e.g. two eyes, pointy ears, whiskers)

OUTPUT

Recap of Image Recognition

INPUT

⭕

An image

➜

LEARNING ALGORITHM

Learn features from training data

➜

OUTPUT

CAT

A label for the image

Let’s test how a machine uses prediction!

Open the Teachable Machine website on your computer
Click Image Project
Click Standard Image Model
Add a “class” so there are 3 classes — Rock, Paper, and Scissors
Use your webcam to take lots of photos of your hand shaped like each class

Take photos against a plain backdrop! This is very important for the machine to learn correctly!
Upload the pictures from the three groups into the Standard Image Model
Once uploaded, click the ‘Train Model’ button
Once trained, export your model and test it by putting your hand in front of the camera!

Teachable Machine: Dataset → Algorithm → Prediction

Recap of Image Recognition

INPUT

✊

➜

LEARNING ALGORITHM

Rock · 67 samples ✊

Paper · 40 samples ✋

Scissor · 45 samples ✌️

➜

OUTPUT

“Rock!”

AI predictions

Remember that AI makes predictions. AI systems don’t ‘know’ answers. They guess or estimate them!

Instead of being certain, AI often predicts a range of likely outcomes.

AI predicting Cat 0.98 and Fluffball 0.65

Confidence Intervals

A confidence interval describes the range an answer is expected to fall within and how confident that system is about that range!

Output

Paper

47%

Sciss…

Rock

52%

Output

Paper

98%

Sciss…

Rock

Output

Paper

Sciss…

52%

Rock

47%

In this lesson we learnt:

What is AI?
What are some examples of AI?
What are the 5 Big Ideas of AI?
The data-algorithm-prediction model
What is data?

Students! Don’t forget to submit your competition entries

Ready for the next lesson?

Mark this lesson as COMPLETE ✔ and click the “NEXT LESSON” button below.

Teachers! Please remember to submit any feedback on this lesson in the link below.

Adapted from lessons developed by

Personal Robots

Licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

Powered by MIT RAISE · i2Learning

With contribution from CS in Schools · 5 Big Ideas of AI from AI4K12.org