When assessing the regulatory activities of the European Union, it can of course be pointed out that the entire legislative process is running too slowly. However, I have the impression that the proposals for “controlling” AI may finally be successful. Without the implementation of a holistic solution, in which the framework of conduct / ethical rules will be imposed on the creators and the validation of AI systems by the state will be introduced, one cannot speak of any sovereignty. It seems that only such a duopoly can lead to the use of brilliant AI solutions, minimizing the risks associated with it. However, without strong state organs, this process will not be adequately secured. This article proves that sovereignty over AI seems to be a sine qua non condition for us to be safe in the understanding of dominating processes.