AI Ethics & Safety
AI is now used to decide loans, jobs and what people see. When it is wrong, real people are hurt. This lesson shows how bias gets in, how to check for it, and a short list to use before you ship.
New words
- Bias
- A pattern where a system treats some groups worse than others.
- Fairness
- Treating groups in a way that is just. There are several ways to measure it.
- Wrong rejection
- A person who deserved a yes gets a no from the model.
- Consent
- A person clearly agrees to how their data is used.
- Deepfake
- A fake photo, voice or video made by AI that looks real.
- Human in the loop
- A person checks or approves the AI before a risky action.
What you will learn
- The loan that Meera did not get
- AI copies its data, including the unfair parts
- How to check for bias: compare the errors
- Try it: a loan model with two groups
- There is no single fair setting
- Privacy: take less, protect more
- Honesty: say when it is AI
- Safety for the apps you build
- Who is responsible, and what about the law?
- Trade-offs
- Interview tip
- Key takeaways
10 quiz questions at the end.
The full lesson, with interactive demos and the quiz, opens after you sign in with Google. It is free.