AI Ethics, Governance & the EU AI Act training course

The EU AI Act — Foundations & Structure

• Risk-tier classification
• Prohibited practices
• High-risk AI categories
• General-purpose AI (GPAI)
• Enforcement and fines
• Extraterritorial reach

Classifying Your AI Systems

• Risk classification exercise using real and case-study systems
• Case studies across HR, credit, and healthcare
• Group discussion on grey areas and ambiguous classifications

AI Ethics — Principles, Bias & Fairness

• Fairness and non-discrimination
• Types of algorithmic bias
• Transparency and explainability (XAI)
• Human autonomy and oversight
• EU Ethics Guidelines for AI

Bias Audit Simulation

• Demographic parity analysis
• Disparate impact detection
• Documenting findings — no coding required

GDPR, Data Protection & AI

• Automated decision-making (Article 22)
• Lawful basis for AI training data
• Data minimisation vs. AI performance
• DPIAs for AI systems
• Rights of data subjects

Building an AI Governance Framework

• AI register and inventory
• Roles: AI owner, deployer, provider
• Model cards and system cards
• Incident response for AI
• Procurement checklists
• Third-party and supplier obligations

Drafting AI Conformity Documentation

• Technical documentation requirements
• Conformity assessment process
• CE marking for AI
• Post-market monitoring plan

Transparency, Explainability & Human Oversight

• Explainability techniques (LIME, SHAP)
• Transparency obligations under the Act
• Human-in-the-loop design
• Logging and audit trails
• Right to explanation in practice

Governance Gap Analysis

• Gap analysis template
• Risk prioritisation matrix
• Remediation roadmap

Capstone: End-to-End Compliance Challenge

• Team-based real-world scenario
• Risk classification
• Governance artefact creation
• Presentation and critique