For Who?
Everyone who wants to know how to practicallyimplement the EU AI Act and/or is willing to take up the role of AI Complianceofficer (AICO). No previous knowledge is required.
Goal of the training?
Understand how to implement the AI Act and beable to start taking up your role as AI Compliance Officer.
What is the training about?
Participants will obtain practical knowledgeconcerning:
- Methodologies and tools toidentify and classify AI systems,
- Designing and implementing anAI Governance Framework (as part of the EU AI Act’s Quality ManagementSystem),
- Developing TechnicalDocumentation,
- Defining & managing AIrisks in a risk mgt. system and as part of a risk appetiteframework,
- Delivering and maintainingconformity assessments,
- Handling self-certificationincluding the end-to-end CE labelling process,
- Fulfilling maintenance taskssuch as human oversight, drift & bias monitoring, incident handlingetc.
Training Approach
A single practical AI use case will be coveredthroughout all training modules, allowing participants to understand how eachrequirement is applied in real life.
The Quality Management System and TechnicalDocumentation will serve as the foundation – further enriched with riskmanagement, data governance, information security, FRIA’s,... conformityassessments, (self-) certification and EU labeling, maintenance etc.
The training will distinguish between thedifferent organizational roles, different AI risk classifications and links toother applicable regulatory obligations (like GDPR) while making sure it fitsevery industry and both the public and private sector.
This training will be followed by a set ofspecific ½ day trainings covering the Quality Mgt. System & CEN Norm,Technical documentation (& CEN norm), Risk mgt. - AI risk ontologies &Risk mgt. systems for AI in practice, Data governance - Data Quality – DataOrigination and Bias prevention & remediation, Information Security andFundamental rights impact assessments.
Training Agenda
| # | Module | What we cover |
|---|
| 1 |
EU AI Act Fundamentals & Regulatory Landscape |
- EU AI Act: scope, objectives & key regulatory requirements
- EU AI Act vs GDPR & other applicable regulations
- Market Surveillance Authorities & their role
- The role of the AI Compliance Officer
- AI value chain: provider, deployer & other actors
|
| 2 |
AI System Identification, Classification & Regulatory Obligations |
- AI system identification checklist & decision trees
- AI inventory, registration and ownership
- AI risk classification & applicable regulatory obligations per actor
- Compliance processes, responsibilities & RACI
|
| 3 |
Implementation Methodology |
- The “pyramid” methodology for high-risk AI systems
- Overview of all requirements and their dependencies
- Practical implementation roadmap
|
| 4 |
Starting from the Top: the Conformity Assessment |
- Conformity assessment details and practical requirements
- Use and positioning of the conformity assessment as part of the certification process
|
| 5 |
Building the AI Quality Management System (QMS) |
- The nature and all aspects of the QMS (incl. AI governance, AI lifecycle management, data preparation, ...)
- Link to an AI governance framework, regulatory strategy, policies & procedures
- Documenting and keeping track of the QMS
- Integration with existing organisational processes
- The official CEN norm on QMS
|
| 6 |
AI Technical Documentation & Evidence Management |
- Technical documentation explained (general description, detailed description, monitoring, functioning & control of an AI system)
- Appropriateness of performance metrics, link to the risk management system, relevant changes & harmonised standards (if published)
- Technical documentation in relation to the declaration of conformity
- Evaluation of post-market technical documentation
- Relevant norms
|
| 7 |
AI Risk Management System & Enterprise Risk Assessment |
- Risk management system & enterprise risk management, and their central role in the EU AI Act
- AI risks vs a traditional risk taxonomy and risk appetite framework (incl. KRIs)
- Deep dive into specific AI risks such as bias, representation, hallucinations, excessive agency, ...
- Risk assessment methodology and application: inherent risk, controls, mitigation & residual risk
- Risk & control matrices, monitoring, reporting and escalation in practice
- Link with post-market monitoring and incident management
|
| 8 |
Fundamental Rights Impact Assessment (FRIA) |
- FRIA requirements under Article 27 & applicability
- Identifying and evaluating fundamental rights in practice, incl. different FRIA templates
- Mitigation measures, documentation, monitoring and updates
|
| 9 |
High-Risk AI System Operational Requirements & Controls |
- Data governance, data quality & technical requirements
- Accuracy, robustness, data protection, cybersecurity & testing
- Transparency, human oversight & record-keeping
- AI literacy, provider/deployer obligations & third-party AI governance
|
| 10 |
From Conformity Assessment to Certification and Labelling |
- Key requirements and supporting documentation
- Certification & information obligation
- (Self-)assessment, CE marking & registration, where applicable
|
| 11 |
AI Lifecycle Management & Post-Market Controls |
- AI lifecycle governance & control gates
- Post-market monitoring & performance oversight
- Incident, change & reassessment management
- Human oversight across the entire lifecycle
- AI inventory, registration, QMS & technical documentation maintenance
|
| 12 |
Final Compliance Demonstration & Practical Assessment |
- End-to-end case study: summary of all deliverables
- Evidence traceability & audit readiness
|
