The AI Engineer's Guide to Surviving the EU AI Act

Chapter 5. AI Engineering for High-Risk AI Systems

This chapter explores the comprehensive requirements for high-risk AI systems and how to meet them proactively by implementing AI engineering practices to ensure compliance with the EU AI Act (see Figure 5-1 for a visual of the steps to take to move toward compliance). In the previous chapter, I introduced the foundational steps for compliance: creating an inventory of current and planned AI systems, identifying the risk category of each, and determining whether your organization acts as a provider or deployer of these systems. These steps are essential for understanding which systems fall under the high-risk category and what compliance obligations need to be addressed.

Figure 5-1. This chapter focuses on the requirements for high-risk AI systems and the operationalization of compliance for such systems. See Chapter 1 for an explanation of the end-to-end process steps toward EU AI Act compliance.

Two guiding questions to consider in this chapter are:

• To comply with the EU AI Act, what requirements must high-risk AI systems fulfill?

• What processes, structures, and AI engineering practices need to be established to comply with the EU AI Act?

Let’s start by exploring the requirements for high-risk AI systems, translating them into AI engineering practices, and defining the notion of AI engineering for the EU AI Act in this ...