Human Where?

Abstract


 
 
 We undertake an expansive examination of the terms Human-in-the-loop, Human-on-the-loop, Human-out-of- the-loop, and Human-in-command, as used recently in AI development, relative to their ethical implications and implicit assumptions. Tracing the history and development of the ‘Human ...’ terms, we explore the contexts and uses present from their beginnings. We follow with a discussion of the ethical outlook which the origins of the terms and their recent rebranding for AI development under the notion of oversight have engendered. Drawing upon certain insights of Bruno Latour for support, we suggest that Latour’s ‘forgotten ethical intermediaries’, folded into our technologies, have their analogue in the view of the human as a component of automated systems alternating with a role of human oversight. We argue that a more ethical human relation to technology can be recovered through an expansive emphasis on human participation in technology producing communities. Finally, we present a flexible new scale, the IGP scale, to rate such participation.