In the framework of the 1st TAILOR Conference, Hildebrandt will discuss the relationship between the 7 key requirements for trustworthy AI of the HLEG on AI and the proposed AI Act. Her first point will be that for these requirements to become real we need to turn them into legal obligations with legal effect. She will focus on some of the core legal requirements for high risk AI systems that seem pivotal for TAILOR: the risk management system, the quality management system, data governance, human oversight and accuracy, robustness and cybersecurity. Her second point will be that many of these requirements are prime examples of legal protection by design; AI systems should be tailored in ways that combine functionality with practical and effective protection against risks to safety, health and fundamental rights.

Foundation of Trustworthy AI: Integrating Learning, Optimisation and Reasoning (TAILOR) is an EU-funded ICT-48 Network aimed building the capacity of providing the scientific foundations for Trustworthy AI in Europe by developing a network of research excellence centres leveraging and combining learning, optimization and reasoning.

For slides, click here.