Hoo and Ray Go Shopping

The lesson begins with a story in which the robots Hoo and Ray go shopping with the task of bringing back apples and pineapples. Even though they have expert books full of information, they’re unable to find the fruit — because they expect to be looking for apple trees or a bush full of pineapples. The teacher leads a discussion about how people form mental models based on varied experiences and examples, whereas machines learn only from the data we give them. Children reflect on the many different forms an apple or pineapple can take — and how it might be difficult to recognize them if a model wasn’t trained on a diverse enough image set. In a hands-on activity, students train their own machine learning model using Teachable Machine — they prepare a training set and then test how well their model recognizes different objects. This is followed by a discussion about how bias arises and how it can affect AI system decisions — for example, in facial recognition or recommendation algorithms. The lesson ends with a reflection, where students share their insights and discuss how machine learning models could be improved by providing higher-quality and more diverse data.