P.B.McIlvenny
English Department, University of Oulu, PL 191, SF-90101 Oulu, Finland
SYNOPSIS
Unfortunately, the early work on machine intelligence has somewhat hindered the legitimate application of the social sciences to the artificial intelligence debate. Indicative of the narrowness resulting from this exclusion is the discourse of the Turing test, which embodies an undesirable set of assumptions about intelligence and promotes an asocial computer. So as to step out of the traditional concern with demarcating what it is to be human as opposed to an animal or machine, I consider if we can we say a priori that machines cannot be social and in what sense one might say that machines are social. In order to examine this issue, I draw upon the social sciences, which supply alternative ways of talking about communication, activity and knowledge within social communities. After considering distributed artificial intelligence and the evidence for natural but marginal forms of social being, I suggest that through the realisation of artificial multiple-entity situated activity systems we can investigate the evolution and development of novel and authentic forms of 'proto-practice' and community.
Keywords
Animal communities, Distributed artificial intelligence, Ethnomethodology, Social machines, Situated activity systems, Turing test discourse