The Computerized Socialbot Turing Test: New Technologies of Noopower

Abstract

Networks of socialbots are spreading across social media sites such as Facebook and Twitter. Socialbots are software systems built to automatically control social media profiles. Importantly, unlike their more benign predecessors, socialbots are built to appear to be human; they respond to direct messages, gain followers, and subtly shape online interaction. As an initial study of socialbots, this paper draws on the heritage of Alan Turing's famous test of computer intelligence. Following Turing's analysis of the human mind, I suggest that social media's culture of confession is producing a massive dataset of the internal states of mind in human beings. Once this codification is done, then a universal Turing machine (in this case, socialbots) can imitate the human in a modern day Turing test. Socialbots' success in this regard certainly raises the specter of machine intelligence. But more importantly, socialbots' success demonstrates that social media capitalism is getting more and more skilled at alienating the fruits of laborers, in this case data appropriated from social media users via surveillance and standardization. Moreover, I ultimately argue that socialbots are new technologies of noopower. Noopower is power over and through attention and memory. Derived as they are from the patterns of social media user activities, and because they can pass as human within social media, socialbots can be key technologies in shaping user thoughts and perceptions.