Course 2 · Unit 4

AI ethics, bias and regulation

4 lessons
≈ 10 h of study
Level: intermediate

Building systems that work for everyone and comply with the law. Bias and fairness metrics, privacy and the GDPR, the EU AI Act and responsible deployment practices (model cards, human oversight, red-teaming and incidents).

Create my account See pricing

Topics covered

Lessons in this unit

Bias and fairness in ML models 65 min
Where a model's biases come from, how they're measured with per-group fairness metrics (demographic parity, equal opportunity, equalized odds, calibration) and why they can't all be satisfied at once.
Privacy and data protection in AI 60 min
The essentials of the GDPR for people who build AI systems —personal data, lawful bases, minimisation, DPIA—, the difference between anonymising and pseudonymising, and how a model can leak its training data.
The EU Artificial Intelligence Act (AI Act) 60 min
How Regulation (EU) 2024/1689 classifies AI systems by risk, what obligations providers and deployers have, what's required of general-purpose models and when each part applies.
Responsible deployment — document, oversee, attack and respond 55 min
Model cards and datasheets, human oversight that actually works, red-teaming before launch, and an incident response process for when something goes wrong.

Prerequisites

Before this unit it helps to have done:

Machine Learning (Course 2 · Unit 3)

The full explanations, auto-graded exercises, exams and projects are inside the academy.

Every lesson you complete gives you 10 yang, the academy's currency, and every unit exam you pass, 50.

Shall we start?

Create your account and activate your subscription: you get the whole syllabus from day one.