Nowadays, AI-based systems are becoming increasingly popular in the automotive industry (e. g. driving assistance systems, autonomous driving functions etc.). However, appropriate standards, methodologies and requirements to cover the AI-specific risks are still missing. The aim of this project is to develop generic requirements, corresponding test methods and tools to assess such risks of AI-based systems. Due to the complexity and the wide range of applications, the scope of this work is limited to certain aspects in the automotive context.