What is Regulating AI Systems

Frameworks for the regulation of artificial intelligence (AI) systems are emerging; some are based on regulation theories; others are more technologically focused. Regulation of AI systems is likely to emerge in an ad-hoc, unstructured, and uncoordinated fashion that renders high level frameworks philosophically interesting but of limited benefit in practice. In this paper, the task of arriving at a collection of interventions that regulate an AI system is taken to be a process-oriented problem. It presents a process-oriented framework for the design of regulating systems by deliberating groups. It also discusses regulations of AI systems and responsibility, mechanisms and institutions, key elements for regulating AI systems. The proposed approach might facilitate research and development of responsible AI, explainable AI, and ethical AI for an ethical and inclusive digitized society. It also has implications for the development of e-business, e-services, and e-society.