Safety
AI Driver's Licence
Five lessons on how AI works, where it appears and how to use it safely. Ends with a digital badge.
How to bring AI projects into the classroom through the Curriculum Connections
ACARA's Artificial Intelligence Curriculum Connection shows where AI already sits in the Australian Curriculum Version 9. Here is how to turn it into a project your class can run this term.
What the Connection says
Start with the three key aspects.
The Connection guides teachers to content already in the curriculum. The explicit content sits in Mathematics and Technologies.
Understanding how AI works
Data, algorithms, pattern recognition, chance and data quality. It stresses that generative AI is not a search engine.
Types of AI (digital tools and AI systems)
How systems are designed and used, and how to tell whether a digital tool has AI in it.
Responsible use and application of AI
Fairness, transparency, privacy, bias and harm, and evaluating what an AI system produces.
By year level
What students are asked to do.
Summarised from ACARA's description of the Connection. Check the content descriptions for your year level before planning.
Years 5–6
Explore the output of AI systems, and use critical thinking and research to evaluate what a chatbot says.
Years 7–8
Weigh the advantages and disadvantages of AI. Design and trace algorithms, implement them in a programming language, and work with data, including sampling.
Years 9–10
Model real problems with functions. Design and validate algorithms. Explain how AI systems manage and secure data, and recognise how bias affects machine learning.
Safety
AI Duty of Care
Safe and responsible use of AI at school, for the adults in the building.
Learning
The capstone project studio
Students pick a problem, build a solution, then test it with simulation and consequence testing.
Learning
ML labs
Students train models and explore real datasets in the browser.
Where chatbots and agents fit
Don't just use one. Build one.
Using a chatbot or an agent does not, on its own, meet a content description. Building a small one puts the data, the algorithm and the evaluation in the students' hands, which is where the curriculum content sits.
Instead of using a chatbot
Let's build a small language model.
Students train a tiny model on text they choose and watch it guess the next word. Then they find out why it sounds confident when it is wrong.
- Collect the text and split it into tokens
- Train the model to predict the next word
- Generate, then test where it makes things up
Small language model · next word
The students opened the ▍
door
book
laptop
volcano
An agent · the phases of problem solving
1Define
What is the goal, and how will we know it worked?
2Break down
Split the goal into steps small enough to check.
3Choose tools
Decide what the agent is allowed to use, and what it is not.
4Act and observe
Run one step, then read what came back.
5Check
Did it work? If not, find the step that failed.
Instead of using an AI agent
Let's build one instead.
An agent is problem solving written down: a goal, a plan, some tools and a check. Students build a simple one, one phase at a time, and learn the phases by having to spell each one out.
The agent is the vehicle. What students keep is the habit of defining a problem, breaking it down and checking the result.
Learning · courses for Years 7 to 12
Stronger tools. Tougher problems.
Source: ACARA, Curriculum connection: Artificial intelligence (AI), Australian Curriculum Version 9. Australian Curriculum material is © ACARA 2010 to present, licensed under CC BY 4.0. ACARA does not endorse any product that uses the Australian Curriculum. The course concepts and the view on chatbots and agents are ML4Schools' own, not ACARA's.