Learn the fundamentals of how to adapt to and comply with AI regulation in this introductory course on auditing AI systems for bias and discrimination.
Overview
Syllabus
Introduction
- Welcome to the new world of AI audits
- What is an AI audit?
- How are audits used?
- The state of AI legislation
- Ethics of scoring and classifying humans
- AI audit limitations and opportunities
- Development workflows
- AI performance
- Statistical parity
- Data for auditing AI
- Sources of bias in data
- Types of bias and data sampling methods
- Why explainability matters
- Levels of transparency
- Responsible AI principles: Compliance
- Preparing for AI regulation
- Types of model audits
- Stages of a model audit
- Model audit: Home loans
- Auditing training data
- Audit outcomes: Explainability statements
- Continuous audits
- Generative AI
- Next steps
Taught by
Ayodele Odubela
