Computational Neural Modeling and the Philosophy of Ethics

Abstract

Introduction T here are different reasons why someone might be interested in using a computer to model one or more dimensions of ethical classification, reasoning, discourse, or action. One reason is to build into machines the requisite level of “ethical sensitivity” for interacting with human beings. Robots in elder care, nannybots, autonomous combat systems for the military – these are just a few of the systems that researchers are considering. In other words, one motivation for doing machine ethics is to support practical applications. A second reason for doing work in machine ethics is to try to better understand ethical reasoning as humans do it. This paper is motivated by the second of the two reasons (which, by the way, need not be construed as mutually exclusive). There has been extensive discussion of the relationship between rules, principles, or standards, on the one hand, and cases on the other. Roughly put, those stressing the importance of the former tend to get labeled generalists, whereas those stressing the importance of the latter tend to get labeled particularists. There are many ways of being a particularist or a generalist. The dispute between philosophers taking up these issues is not a first-order normative dispute about ethical issues. Rather, it is a second-order dispute about how best to understand and engage in ethical reasoning. In short, it is a dispute in the philosophy of ethics.