This book offers an interdisciplinary interrogation of the four concepts of fairness, accuracy, knowability and transparency that have long been recognised as cardinal in the regulation of Artificial Intelligence (AI).
In the EU’s AI Act, fairness and transparency are integrated in safety standards tailored for high-risk AI systems, with the purpose of enhancing the auditability, knowability and explainability of those systems’ output. However, practice has shown that the scrupulous observation of those standards is not invariably conducive to accurate and fair AI outcomes. It also does not systematically facilitate the access to knowledge about possible rights violations. Focusing on the concepts of fairness, accuracy, knowability and transparency, the book explores the multifaceted nature of their interrelationship in relation to: the complementarity and tension between them; their operationalisation through regulation, compliance and adjudication; and their expression in the form of individual entitlements. With twelve chapters authored by experts in law and political science, the book combines epistemology and ethics in order to offer unique insights into the foundational concepts that inform both the theory and practice of AI regulation.
This book will be of interest to legal scholars, as well as practitioners and policy-makers, with interests in the ethics and regulation of AI.
