Course / Domain III
Governing AI development
The controls and records that belong at every stage when you build an AI system: design, data, training, testing, release and monitoring, plus generative AI specifics.
21–25 scored items0/6 lessons done
Lesson 3.1 · 12 min
Plan and design: scoping the system
- Run a use-case intake and risk classification
- Choose the right impact assessment
- Define success, fairness and oversight requirements up front
Lesson 3.2 · 14 min
Data governance and bias
- Apply data provenance, rights and quality checks
- Identify sources of bias
- Explain why fairness metrics conflict
Lesson 3.3 · 9 min
Building and training with privacy in mind
- Select privacy-enhancing technologies
- Explain model documentation and versioning
Lesson 3.4 · 13 min
Testing, explainability and security
- Plan a test program covering performance, fairness, robustness and security
- Distinguish global and local explanations
Lesson 3.5 · 10 min
Release, monitoring and change control
- Design a release gate
- Plan post-deployment monitoring and change management
Lesson 3.6 · 12 min
Governing generative AI
- Identify risks specific to generative AI
- Select controls for generative systems