Paper Short Abstract:

Drawing on 2 years of fieldwork in a Japanese robotics laboratory, I analyse the recent 'predictive turn' in robotics, AI and neuroscience. In so doing, I unearth the vernacular specifications of surprise and futurity stemming from it, and speculate upon how these visions reshape humanity.

Paper Abstract:

This paper draws from 2 years of fieldwork in a Japanese robotics laboratory, and focuses on the specific sets of theories and experiments emerging at the conjunction between AI, robotics and neuroscience, attempting to explain, model, and ultimately reproduce living processes.

Particularly, such endeavours focus on modelling life as a struggle to minimise one's 'surprise' in regards to environmental stimuli, optimising one's effort by learning to predict the structures of the world in increasingly accurate terms.

Starting from these premises of life as a predictive machine, I first explore what such a conceptualisation does to our understanding of the (all too human) experiences of gaining information, and being surprised. Further, I analyse a specific experiment carried out in the laboratory which operationalises such an understanding. I highlight how such a vernacular re-framing might be radically different from anthropological modes of thinking about futurity, rather showcasing what happens when prediction becomes a tool to control information and surprise. Lastly, I report a thought experiment emerged in conversation with my informants: what would happen when a robot is simply unable to be surprised anymore?

Starting from such impossible speculations which stretch the conceptual limits of these theories, I trace the different ways in which we can think about prediction, the predictive capacities of living beings, and the opportunities afforded by a predictive anthropology. In doing so, I draw on Derrida's idea of 'absolute surprise' to retrace the contours of a mode of predicting that creates ever-so-new potentials, rather than drying up possibilities.