To the extent there is any consensus among States, ban advocates, and ban skeptics regarding the regulation of autonomous weapon systems (AWS), it is grounded in the idea that all weaponry should be subject to "meaningful human control." This "intuitively appealing" principle is immensely popular, and numerous States have explicitly declared their support for it or questioned the lawfulness of weapons that operate without such control. Lack of opposition has led some to conclude that it is either a newly developed customary norm or a preexisting, recently exposed rule of customary international law, already binding on all States.
But this broad support comes at a familiar legislative cost; there is no consensus as to what "meaningful human control" actually requires. State X might define meaningful human control to require informed human approval of each possible action of a given weapon system (maintaining a human being "in the loop"); State Y might understand it as the ability of a human operator to oversee and veto a weapon system‘s actions (having a human being "on the loop"); and State Z might view the original programming alone as providing sufficiently meaningful human control (allowing human beings to be "off the loop"). As the Czech Republic noted, in voicing its belief that "the decision to end somebody‘s life must remain under meaningful human control, . . . [t]he challenging part is to establish what precisely 'meaningful human control' would entail."
This paper describes attempts to clarify what factors are relevant to meaningful human control, discusses benefits associated with retaining imprecision in a standard intended to regulate new technology through international consensus, and argues that the standard‘s vagueness should be limited by an interpretive floor. "Meaningful human control" as a regulatory concept can usefully augment existing humanitarian norms governing targeting—namely, that all attacks meet the treaty and customary international law requirements of distinction, proportionality, and feasible precautions. However, it should not be interpreted to conflict with these norms nor be prioritized in a way that undermines existing humanitarian protections. [..]
Rebecca Crootof, A Meaningful Floor for "Meaningful Human Control", 30 Temp. Int'l & Comp. L.J. 53 (2016).