Week 11 · 45-60 min

Ethics and Bias in AI

Week Summary

A quick preview of the main ideas in this week before you begin the slides.

  • Where Bias Comes From in AI
  • Real-World Examples of AI Bias
  • Why Representative Data Matters
  • Transparency and Explainability

Glossary

Review these key terms before starting this week. They are kept simple so students can learn the new vocabulary before moving into the slides.

Where Bias Comes From in AI
Where Bias Comes From in AI
Real-World Examples of AI Bias
Real-World Examples of AI Bias
Why Representative Data Matters
Why Representative Data Matters
Slide 1💡

Where Bias Comes From in AI

Learning goal

Understand how bias enters AI systems and why ethics matter.

Read one card at a time. Focus on the main idea first, then look at the visual on the right to help it stick.

Bias often begins in biased training data.

Historical inequality can be reflected in dataset patterns.

Model design choices can also introduce bias.

Bias may be unintentional but still harmful.

Recognizing sources is the first step toward addressing them.

Remember

Bias in AI often starts before training, with the data and assumptions that shape the system.

Where Bias Comes From in AI

Where Bias Comes From in AI

Slide 1 of 8

Use arrows or dots to move through the lesson

Practice

  1. Write 2-3 sentences about what you learned this week.

Coming Next Week: Next: Review and Wrap-Up

Project Brief

Complete the "Train and Evaluate a Simple Model" project across weeks 11-12.

What to Build

  1. Day 1: Choose or receive a small dataset and define the prediction task and question.
  2. Day 2: Clean and structure the dataset; identify features and label.
  3. Day 3: Split data into training and testing sets.
  4. Day 4: Choose and train an appropriate model such as linear regression, logistic regression, or a decision tree.
  5. Day 5: Generate predictions on the test data.
  6. Day 6: Calculate appropriate evaluation metrics such as accuracy or error and a confusion matrix if relevant.
  7. Day 7: Compare training and testing performance to look for overfitting or underfitting.
  8. Day 8: Write a short reflection on potential bias or limitations.
  9. Day 9: Prepare a short presentation summarizing the task, model, results, and limitations.
  10. Day 10: Present findings and answer questions.

Self-Assessment Rubric

  • Correct ML workflow implementation — 30%
  • Appropriate model choice and evaluation — 25%
  • Depth of ethics, bias, and limitations reflection — 20%
  • Code clarity and comments — 10%
  • Presentation clarity — 15%