Data protection in the AI age

Data protection aspects of prohibited AI practices and the use of certain AI systems generating risks

Abstract

The widespread use of AI today poses particular challenges for the protection of personal data. Thus, through the use of AI systems, the behavior of users can be easily observed, their habits and character can be mapped, from which AI can draw almost innumerable conclusions. The European AI Regulation has defined the scope of AI practices that can be considered prohibited in democratic societies with regard to the risks arising from excessive surveillance and certain other harmful AI practices. In addition, there are other AI applications that can be considered high-risk or potentially harmful and risky, and which have therefore attracted considerable attention in the context of data protection and with respect to the related risks to privacy. In line with the above, the aim of this paper is to present the data protection aspects and main risks associated with the use of prohibited AI practices and certain AI systems, including the challenges of biometric identification, certain risk assessment systems and web scraping.