In Search of Trustworthy AI: Can the Law Help?

In recent years, policymakers and computer scientists have increasingly embarked upon a quest for “trustworthy AI”. However, many challenges such as opacity and discrimination continue to haunt AI, and in particular deep learning. This talk discusses some key possibilities the law offers to tackle these challenges. More specifically, the talk discussed what resources the General Data Protection Regulation (GDPR), the proposal for the EU AI Act and the EU Digital Markets Act hold to set incentives for explainable AI and algorithmic fairness. The talk concludes with some policy options to further address these concerns, both on a legal and a technical level.