A Declarative Modular Framework for Representing and Applying Ethical Principles
Fiona Berreby, Gauvain Bourgne, Jean‐Gabriel Ganascia · 2017
This paper investigates the use of high-level action languagesfor designing ethical autonomous agents. It proposes a noveland modular logic-based framework for representing and rea-soning over a variety of ethical theories, based on a modi-fied version of the Event Calculus and implemented in An-swer Set Programming. The ethical decision-making processis conceived of as a multi-step procedure captured by fourtypes of interdependent models which allow the agent to as-sess its environment, reason over its accountability and makeethically informed choices. The overarching ambition of thepresented research is twofold. First, to allow the systematicrepresentation of an unbounded number of ethical reason-ing processes, through a framework that is adaptable andextensible by virtue of its designed hierarchisation and stan-dard syntax. Second, to avoid the pitfall of much research incurrent computational ethics that too readily embed moralinformation within computational engines, thereby feedingagents with atomic answers that fail to truly represent un-derlying dynamics. We aim instead to comprehensively dis-place the burden of moral reasoning from the programmerto the program itself.