Ethical rules for robots require human specifications of their key variables

Evidence

As roboticist Noel Sharkey (who is fiercely opposed to the development of such programs) explains, the deliberation software for an “ethical robot” must, like any other program, necessarily incorporate not only rules but also specifications. In other words, an order to “target only legitimate targets” is an empty operation if there is no specification of what is covered by the variable term “target.” Similarly, one can attempt to codify the principle of proportionality (good luck with that), but it will always be necessary to specify, whether directly or indirectly, what constitutes the acceptable proportion of the number of civilians killed to the military advantage expected. In short, the parameters of the decision need to be specified, and that specification is not done by the program itself. A higher-level choice is needed—a decision about the decision. — Grégoire Chamayou, A Theory of the Drone, 676

Categorization

• DEFINE: Introduces the concept of a ‘decision about the decision’ — the human specification that precedes any automated rule.
• XREF: Connects to the reader’s interest in how centralized command moves into programmatic specifications rather than verbal orders. — connects to: Embedding ethics in software cannot guarantee lawful conduct because software is reprogrammable.; Rule-based machine ethics produce unintended emergent consequences — , The Braided Reading EPUB: 5-Year…, 1591; Automation requires active human work to organize the relinquishing of subjectivity — Grégoire Chamayou, A Theory of the Drone, 650 (+9 more)

Related

• Correct action requires coding rules, matching repertoires, and sufficient information. (similarity 0.91)
• Embedding ethics in software cannot guarantee lawful conduct because software is reprogrammable. (similarity 0.86)
• Rule-based machine ethics produce unintended emergent consequences (similarity 0.84)
• Automation requires active human work to organize the relinquishing of subjectivity (similarity 0.82)
• Machines could learn human values by observing behavior instead of being programmed (similarity 0.80)
• The danger of military robots is their incapacity to disobey (similarity 0.80)
• AI alignment may require machines to pursue human objectives, not their own (similarity 0.79)
• The concept of Asimov’s robot laws requires reevaluation in modern contexts. (similarity 0.78)

Broader Connections

• Machines could learn human values by observing behavior instead of being programmed #contradicts (0.70) — Brian Christian, The Alignment Problem, 1086
• AI researchers aim to align general-purpose systems with hard-to-specify human intentions #corroborates (0.70)
• Ensuring AI captures human intent is now a central computer science problem #expands (0.60) — Brian Christian, The Alignment Problem, 65
• Moral rules cannot be created or altered by deliberate human enactment #contradicts (0.60) — HLA Hart, The Concept of Law (Clarendon Law…, 443