EETAS: A Process for Examining Ethical Trade-Offs in Autonomous Systems

Abstract

Public-facing autonomous systems present society with significant ethical challenges, not least of which is the need for stakeholder understanding and discussion of how these systems balance competing ethical principles. In this paper we present EETAS: a structured, gamified process for obtaining stakeholder input into the ethical balances and trade-offs which they consider it acceptable for a proposed autonomous system to make. We describe how outcomes from the EETAS process can be used to inform the design of specified autonomous systems, as well as how the process itself can improve stakeholder engagement and public understanding of ethics in AI and autonomous systems. In support of this we present the findings from an initial EETAS pilot study workshop, which shows an indicative trend of improvement in public understanding and engagement with AI following participation.