Ch 25: AI Governance & Ethics - Introduction¶
Track: Advanced | Try code in Playground | Back to chapter overview
Read online or run locally
You can read this content here on the web. To run the code interactively, either use the Playground or clone the repo and open chapters/chapter-25-ai-governance-and-ethics/notebooks/01_risk_classification.ipynb in Jupyter.
Chapter 25: AI Governance & Ethics — Notebook 01 (Risk Classification)¶
Governance begins by asking how risky a system is. We implement a transparent, rule-based risk classifier and a likelihood × severity matrix.
What you'll learn¶
| Topic | Section |
|---|---|
| The EU AI Act risk tiers | §1 |
| A rule-based classifier | §2 |
| Likelihood × severity matrices | §3 |
Time estimate: 2 hours
Key concepts¶
- Risk tiers — minimal, limited, high, unacceptable.
- Transparency — governance rules must be explainable, not black boxes.
- High-risk domains — employment, credit, biometrics, law enforcement.
- Risk matrix — combine probability and impact into one level.
A transparent, rule-based risk classifier assigns each system a regulatory tier, and a likelihood × severity matrix prioritizes mitigations. Governance starts with knowing how risky a system is.
Run the full notebook in the chapter folder for code and outputs.
Generated by Berta AI