🟣 Year 6 Artificial Intelligence Activity

🤖

AI Prediction Lab

Train an AI, test its predictions and investigate what happens when the training data is incomplete, incorrect or unbalanced.

🧠 You are the AI Trainer!

Modern AI systems can be trained using examples. They look for patterns in the data and use those patterns to make predictions about new information.

📊 DATA

🔍 PATTERN

🔮 PREDICTION

✓ CHECK

In this challenge you will first train the AI. Then you will investigate the quality of its training data.

Year 6 Challenge:
Remember that an AI prediction is not automatically a fact. The quality of the prediction depends on the information the AI has learned from.

Training 1 of 8

Score: 0

Train the AI

Study the labelled examples and identify the pattern.

New data has arrived!

?

Which category should the AI predict?

🧠

Training Complete!

You have just used the basic process behind a machine-learning prediction.

📊
DATA

🔍
PATTERN

🔮
PREDICTION

✓
CHECK

But there is a problem. What if the training data is poor?

A computer does not automatically know that its training data is accurate, complete or fair.

Investigation 1 of 6

Score: 0

🔬 Investigate the Training Data

Look carefully. Is this training data reliable enough for the AI to learn a useful pattern?

Training Data

🏆

AI Prediction Lab Complete!

🤫 What have you discovered?

📊
AI learns from
DATA

🔍
AI finds
PATTERNS

🔮
AI makes
PREDICTIONS

✓
Humans should
CHECK

Important idea:

An AI prediction is not the same as a fact.

AI systems learn from data. If the data is incomplete, incorrect, unbalanced or labelled badly, the patterns the AI learns may also be unreliable.

🚨 Think like a digital detective!

When an AI gives you an answer, do not automatically assume that it is correct.

🧠 Question the information. Check the evidence. Think about where the data came from.