Stage 2: AI literacy for ages 8 to 10

1. Different Kinds of Smart

I can compare what people and machines each do well and pick the right mix for a job, with a reason.

Different Kinds of Smart · The Right Team for the Mission

2. See It, Think It, Check It

I can separate what I observe from what I infer, and check my idea with evidence before I decide.

See It, Think It, Check It · The Evidence Detective Case

3. Questions That Open Thinking

I can turn a vague question into an open, focused one and judge whether an answer is reliable.

Questions That Open Thinking · The Reliable Answer Challenge

4. Design Starts With the Problem

I can write a problem statement from what users say and choose a design by weighing trade-offs.

Design Starts With the Problem · The Trade-off Design Lab

5. The Plan-Monitor-Improve Cycle

I can plan with checkpoints, monitor progress with a measure, and make one improvement based on evidence.

The Plan-Monitor-Improve Cycle · The Team Improvement Challenge

6. From Question to Data

I can plan which data answers a question, collect it fairly, and show it in an honest graph.

From Question to Data · The Honest Graph Studio

7. Features That Matter

I can choose the features that help sort examples and test a simple classifier on new examples.

Features That Matter · Build and Test a Classifier

8. Patterns That Help Us Predict

I can use a pattern in data to make a prediction, record it before checking, and explain why it might be wrong.

Patterns That Help Us Predict · The Prediction Challenge

9. Learning Examples and Hidden Tests

I can keep test examples hidden from training, test a model fairly, and tell memorising from learning.

Learning Examples and Hidden Tests · The Generalization Lab

10. The Dataset Audit

I can audit a dataset for wrong labels and missing groups, fix them, and check the accuracy for every group.

The Dataset Audit · Redesign for Better Representation

11. From Instructions to Flowcharts

I can turn instructions into a flowchart with a start, steps, a decision and an end, and trace it to check it works.

From Instructions to Flowcharts · Design an Algorithm for a Real Service

12. If, Then, Else

I can write if/then/else rules with clear, measurable conditions and test them on boundary cases and on everyone they affect.

If, Then, Else · Build a Decision System

13. Find the Repetition

I can find the repeating unit in instructions, replace it with a loop, and give every loop a safe stopping rule.

Find the Repetition · Loops With a Stopping Rule

14. The Test Case Laboratory

I can write test cases with inputs and expected outputs, find a bug, fix one thing, and retest everything.

The Test Case Laboratory · Fix, Retest and Protect Working Parts

15. Inside an Intelligent System

I can trace data from sensor to prediction to action, and design a smart system with limits, a fail-safe and a human override.

Inside an Intelligent System · Design a Safe Smart System

16. Predicting the Next Piece

I can explain that generative AI makes likely next pieces from patterns, and verify and cite what it produces.

Predicting the Next Piece · Prompt, Compare, Verify

17. Pause Before You Share

I can use a four-check pause before sharing and investigate whether a picture, sound or video might be made by AI.

Pause Before You Share · The Synthetic Media Investigation

18. Data, Permission and Digital Footprints

I can decide what data an app really needs, explain my digital footprint, and spot phishing warning signs.

Data, Permission and Digital Footprints · The Security and Phishing Challenge

19. Who Is Affected and How?

I can map who an automated decision affects, weigh benefits and harms, and design explanation, appeal and human oversight.

Who Is Affected and How? · Explain, Appeal and Oversee

20. Discover, Define and Design

I can run a full design cycle: discover a real need, define it, design and build, test with users, improve and defend my choices.

Discover, Define and Design · Build, Test, Improve and Defend