Artificial intelligence curriculum for primary and secondary schools
AI in Informatics at Secondary and Secondary School
Bias
Card 03 from the Machine Learning deck
Bias
Artificial intelligence systems learn from data prepared by humans. Poorly prepared, unbalanced data, or lack of data, can cause AI to be biased in some way. Students continue with the Alien Detective Agency activity.
Information about lessons
Subsidies
45 minutes
Vintages or preconceptions
2. primary school level and multi-year grammar schools
Aids
Educator: sets of cards in alien families, projection equipment, presentation.
Students: Computer, laptop or tablet (but not with Android OS) for each student or group.
Building stones
Bias, dataset, machine learning
What pupils learn
Machine learning systems are biased.
Bias can be prevented by testing the model and modifying the dataset.
Why is it taught
Goal for the Machine Learning deck of cards: critically assess the decision making of artificial intelligence systems.
How do we know if they've learned
Explain the concept of bias.
Outputs of the RVP
Data, Information, and Modeling
I-9-1-04 evaluates whether the model contains all the data needed to solve the problem; locates the error in the model and corrects it.
Bloom's taxonomy
Understanding: understand how insufficient or poorly representative data leads to bias in AI models.
Application: use the Teachable Machine tool to train and test the model, identify instances of bias in data.
Digital competences
Information and communication
Five Big Ideas
3-C-III Datasets (bias)
