The criminalisation of deep fakes (with particular attention to the implications of the European regulation on Digital Services and Artificial Intelligence)

Abstract

Artificial Intelligence presents risks that are being addressed by different enforcement techniques. A subspecies of these systems, the generative models, enable the creation of all kinds of content, including deep fakes, representations of people performing behaviors that did not take place in the first place. Recently, the criminalization of deep fakes has been raised in both EU and national legal systems. This paper analyzes the response that the aforementioned reforms present to sexual deep fakes, framing them in the complex European regulatory framework of deep fakes, that establishes additional obligations to the operators of these systems relevant to criminal law.