Ethics and agent-based computational social models: Do implementations of social simulations have moral significance?

Abstract

The focus of social simulation on representing the social world suggests an investigation of whether its implementations are inherently value-laden. The purpose of algorithms is both functional and representational. Some algorithms may be said to have functional purposes, such as deleting an email or ordering a list. The purpose in some simulations may be said to be essentially representational – to represent other things in the world, such as a model to investigate the problem of ethnical segregation, as in Schelling’s (1971). Algorithms may also be purposely designed for certain values, according to certain requirements, such as protecting or undermining privacy. Despite articulating values, algorithms need not be essentially value-laden, that is, they need not comprise essential value judgements. In general, it can be said that an algorithm comprises an essential value judgement if, for the same purpose and everything else being equal, designers who accept different value judgements would have rational reasons for designing different algorithms (Kraemer et al., 2011).