Exploring Article 14 of the EU AI proposal: accountability challenges of the human in the loop when supervising high-risk AI systems in public administration

Abstract

This research paper utilises a theoretical approach to analyse Article 14, human oversight, of the proposal for Regulation on the European Parliament and of the Council Laying Down Harmonised Rules on Artificial Intelligence (Artificial Intelligence Act) and Amending Certain Union Legislative Acts and test it against the various challenges of public service organisations (also referred as street-level bureaucracies). This research was chosen because so much emphasis has been given to deploying safe or trustworthy Artificial Intelligence (AI) systems from a “machine design” perspective. However, there is little debate about the effectiveness of other safeguards, such as the human in the loop, to promote safe or trustworthy AI. Therefore, this paper explores the accountability challenges in public administration when adopting the human in the loop as a safeguard to oversee the implementation and proper functioning of high-risk AI systems in different government agencies.

Keywords: Accountability, public service, human oversight, decision-making, street-level bureaucracies, high-risk AI systems, Article 14 EU AI Act Proposal.