Lesson 3: Bias in bias out

In this lesson, students will create their own machine learning model to classify images of apples and tomatoes. They will discover that a limited data set can lead to a flawed ML model. They will then explore how bias can appear in a data set, leading to ML models producing biased predictions.

Learning objectives

  • Describe the impact of data on the accuracy of a machine learning (ML) model
  • Explain the need for both training and test data
  • Explain how bias can influence the predictions generated by an ML model

Key vocabulary

Artificial intelligence (AI), machine learning (ML), supervised learning, classification, training data, test data, accuracy, bias, data bias, societal bias

Lesson structure

  • The three different types of machine learning
  • Supermarket AI application
  • Training a model
  • Bias
  • Student timetable model
  • Reducing bias

Lesson overview

Download lesson resources

To download the lesson plan, slides and activities please log in with or create a free Raspberry Pi Foundation account.
We will then ask you a few questions to help us understand how you will use the resources.