Is the Traditional Method of Regulation (the Legislative Act) Sufficient to Regulate Artificial Intelligence, or Should It Also Be Regulated by an Algorithmic Code?

Abstract

Th e issue of the regulation of artifi cial intelligence (AI) is one of the signifi cant challenges faced by the EU at present. Most researchers focus on the substantive scope of AI regulation, including state law, ethical norms and soft law. In addition to the substantive and legal scope of the regulation, it is worthwhile considering the manner of such regulation.1 Since AI is an algorithmic code, it seems correct to regulate (restrict) AI not so much with traditional law established in natural (human) language as with one implemented into algorithms. Th ey may operate as a tool supporting traditional legislation (RegTech), but it is possible to go further with the issue and create regulation algorithms which implement the law as the eff ective law. However, this requires a new approach to law and legislation – the law as algorithmic code.