Accessible, Cutting-Edge AI Policy Research

On Meaningful Human Control in High-Stakes Machine-Human Partnerships

The autonomy of artificial agents is an important aspect of defining machine-human partnerships. There is no clear consensus definition for the concept of meaningful human control (MHC) (Santoni de Sio et al., 2018). Broadly it connotes that in order for humans to be capable of controlling – and ultimately responsible for – the effects of automated systems, they must be involved in a non-superficial or non-perfunctory way.