Toward a General Logicist Methodology for Engineering Ethically Correct Robots

摘要

It's hard to deny that robots will become increasingly capable and that humans will increasingly exploit these capabilities by deploying them in ethically sensitive environments, such as hospitals, where ethically incorrect robot behavior could have dire consequences for humans. How can we ensure that such robots will always behave in an ethically correct manner? How can we know ahead of time, via rationales expressed clearly in natural language, that their behavior will be constrained specifically by the ethical codes selected by human overseers? In general, one approach is to insist that robots only perform actions that can be proved ethically permissible in a human-selected deontic logic--that is, a logic that formalizes an ethical code. Ethicists themselves work by rendering ethical theories and dilemmas in declarative form and reasoning over this information using informal and formal logic. The authors describe a logicist methodology in general terms, free of any commitment to particular systems, and show it solving a challenge regarding robot behavior in an intensive care unit.This article is part of a special issue on Machine Ethics.