The proposal for the Artificial Intelligence Act is the first comprehensive attempt to legally regulate AI. Not merely because of this pioneering role, the draft has been the subject of controversial debates about whether it uses the right regulatory technique, regarding its scope of application and whether it has sufficient protective effect. Moreover, systematic questions arise as to how the regulation of constantly evolving, dynamic technologies can succeed using the means of the law. The choice of the designation as Artificial Intelligence Act leads to legal-theoretical questions of concept formation as a legal method and legislative technique. This article examines the difficulties of regulating the concept of AI using the scope of the Artificial Intelligence Act as an example.