Distributed mimicry: Big Data and Predictive Analytics

Abstract

This paper argues that mimicry is a central issue in the development of new practices of predictive analytics and big data. The issue concerns the increasingly precise reproduction of human interactional dynamics and their translation to machine and code worlds. Mimicry, in this sense, allows for predictive analytics to simulate a huge range of individual and collective cross-species behaviours. This includes human practices, particularly at the non-conscious, non-verbal level sometimes with an uncanny appearance of intuitive anticipation. The paper takes up different perspectives on mimesis and simulation to discuss the exploitative and emancipatory tensions these articulate. It investigates these through developments in swarm robotics, text mining and recent advances in codifying human synchrony. All of these utilise mimicry as core elements in their development.