Louise Amoore: On Intuition: Machine Learning and Posthuman Ethics

Abstract

Whether in new surgical robots, or in the geopolitics of automated weaponry, drones, and intelligence gathering, machine learning algorithms and operatives are trained for future action via the patterns of ingested past data. What kind of ethics is possible in the context of the intuitive learning of a posthuman composite? Can this form of cognition and action be meaningfully called to account? As Katherine Hayles has written, “what is lethal is not the posthuman as such, but the grafting of the posthuman onto a liberal humanist view of the subject” (1999: 23). Thus, as contemporary legal cases proliferate, they persistently seek an identifiable human subject to call to account – a specific surgeon who made an error, a particular drone pilot or analyst who wrongfully targeted – who is often called the “human in the loop” of semi-supervised machine learning. Yet, machine learning is precisely changing the nature of what it means to be human, so that the errors of a neural net must involve an expanded and distributed sense of ethics. In this paper I propose an alternative mode of ethics capable of responding to the intuitive learning of human and algorithm.