Module 1: Context of the EU AI Act

• Purpose and scope of the EU AI Act
• Primary objectives: smooth functioning of internal market and promotion of human-centered AI
• Scope of application: what is covered and what is excluded
• Key definitions and stakeholder roles (provider, deployer, importer, distributor, etc.)
• Governance structure at EU, national, and operational levels

Module 2: EU AI Act in Depth

• Key provisions and regulatory framework
• Risk-based classification system

  • Unacceptable risk (prohibited AI)
  • High-risk AI systems (requirements and obligations)
  • Limited risk AI (transparency obligations)
  • Minimal or no risk AI
• Requirements for high-risk AI systems

  • Product compliance and CE marking
  • Technical documentation requirements
  • Quality and risk management systems
  • Data governance obligations
  • Conformity assessments
• General-purpose AI (GPAI) models and systemic risks
• Code of practices for GPAI
• Intellectual property implications
• Open-source vs. closed-source considerations
• Provider obligations: conformity assessments, documentation, regulatory notification
• Compliance and enforcement mechanisms
• Penalties for non-compliance
• Accountability and incident reporting

Module 3: Trustworthy AI

• Privacy and data protection under the AI Act
• Importance of transparency in AI systems
• Traceability requirements and implementation
• Data minimization principles
• GDPR principles applied to AI scenarios
• Role of transparency in building public trust

Module 4: Ethical AI

• Key ethical principles in AI development

  • Transparency
  • Accountability
  • Fairness and equality
  • Privacy and data protection
  • Safety and security
  • Human-centric design
• Application of ethical guidelines to real-world scenarios
• Human rights considerations
• Fundamental rights impact assessments (FRIA)
• Importance of human oversight in AI systems
• Human-in-the-loop vs. human-on-the-loop models

Module 5: EU AI Act in Practice

• AI in the public sector

  • Public decision-making
  • Crime prosecution and law enforcement
  • Elections and democratic processes
  • Risks and mitigation strategies
• AI in the private sector

  • Finance and insurance compliance
  • Healthcare applications
  • Employment and education systems
  • Autonomous driving
  • Advertising and tourism
• Sector-specific requirements and best practices
• Impact assessments by industry

Module 6: Frameworks to Support Compliance

• European standards

  • CEN/CLC/TR 18115: Data governance practices
• International standards

  • ISO 42001: AI management systems
  • ISO 23894: AI risk management
  • ISO/IEC TR 24368: Ethical considerations
  • NIST AI Risk Management Framework
• Comparison and alignment of frameworks
• Integration strategies for compliance programs
• Practical application of standards

Lab 1: AI System Classification and Risk Assessment

• Analyze sample AI systems and classify them by risk level
• Identify applicable requirements based on classification
• Document rationale for risk categorization

Lab 2: Fundamental Rights Impact Assessment (FRIA)

• Conduct a FRIA for a high-risk AI system scenario
• Identify affected fundamental rights
• Assess likelihood and severity of impacts
• Develop mitigation strategies

Lab 3: Data Governance Implementation

• Apply Article 10 requirements to a training dataset
• Identify data quality issues and biases
• Implement data governance controls
• Document data lineage and processing activities

Lab 4: Compliance Documentation

• Prepare technical documentation for a high-risk AI system
• Create conformity assessment records
• Develop transparency documentation for users
• Draft incident response procedures

Lab 5: Multi-Stakeholder Compliance Scenario

• Work through a complex scenario involving multiple stakeholders (provider, deployer, importer)
• Assign responsibilities according to the EU AI Act
• Develop compliance workflows
• Create reporting and oversight mechanisms