Building on a structured overview of the regulations of the European AI Act, key aspects will be examined in depth: the role and risk-based classification of AI systems, including high-risk AI, and the resulting obligations. These include requirements for a quality management system, data and model governance, technical documentation, transparency, human oversight, robustness, and monitoring throughout the entire life cycle. Building on this, the interaction with the EU GMP guidelines will be explained – in particular, how requirements from the AI Act can be translated into existing GMP principles and integrated in a practical manner. The focus is on validation, data integrity, a risk-based approach, and change management, supplemented by aspects of traceability and explainability.
Finally, there is a brief outlook on the status of Annex 22 and the possible implications for practical implementation.