The Hollow Human: A Cybernetic Anatomy of AI and Ethics

Peter Asaro began as a historian of science, recovering the Cybernetics of W. Ross Ashby — the intelligence-amplifier, and the mind as an embodied regulator — and helping to salvage Ashby's lost notebooks. He carried that work into a second career campaigning against autonomous weapons, on the principle that the authority to take a human life cannot be delegated to an automated process: meaningful human control. This paper reads the ethics of artificial intelligence through the apparatus Asaro recovered — Ashby's amplifier — joined to Stafford Beer's viable-system model, and turns it on the case his advocacy anticipated. The AI targeting systems used in Israel's assault on Gaza — Lavender, the Gospel, and Where's Daddy? — were not autonomous weapons: a human authorised every strike. Yet meaningful control failed. The failure lies in a configuration the prevailing debate cannot name: the hollow human, an operator present in title and void in function, whose evaluation is reduced by throughput and policy to a formality. The paper's central claim is that the two members of the human–machine coupling fail from a single cause — amplification just is the exceeding of human selective capacity, so to scale the amplifier is to void both the grounding of a target and the checking of it at once, by construction. Grounded in the operators' own testimony and corroborated by four decades of automation research, clinical evidence of automation bias and deskilling, and a controlled simulation, it concludes that meaningful human control is not a policy preference but the amplifier's safety condition — and that a system voiding it to take human lives, at a scale a growing body of legal and scholarly authority characterises as genocide, is for that reason impermissible.