Providers of AI systems must ensure that their products undergo a risk assessment and comply with EU requirements. This includes, among other things, the quality of the datasets used, the traceability of the results, and the provision of detailed documentation. These measures are intended to ensure that AI systems are safe and trustworthy.