The role of data regulation in shaping AI: An overview of challenges and recommendations for SMES

Abstract

In recent debates around the regulation of artificial intelligence, its foundations, being data, are often overlooked. In order for AI to have any success but also for it to become transparent, explainable and auditable where needed, we need to make sure the data regulation and data governance around it is of the highest quality standards in relation to the application domain. One of the challenges is that AI regulation might - and needs to - rely heavily on data regulation, yet data regulation is highly complex. This is both a strategic problem for Europe and a practical problematic: people, institutions, governments and companies might increasingly need and want data for AI, and both will affect each other technically, socially but also regulatory. At the moment, there is an enormous disconnect between regulating AI, because this happens mainly through ethical frameworks, and concrete data regulation. The role of data regulation seems to be largely ignored in the AI ethics debate, Article 22 GDPR being perhaps the only exception. In this chapter, we will provide an overview of current data regulations that serve as inroads to fill this gap.