An Arrow That Steers Itself: A Natural-Law Research Agenda for Governing AI Agents

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

The idea in two sentences AI agents need more than good goals. They need a rule that keeps what they do, the authority to do it, and the purpose it serves bound together from the first step to the last. Summary phrase: An action succeeds morally only when what is done, the authority to do it, and the purpose for which it is done remain ordered together through execution. Key rule: An agent must do a permissible thing, under authority it still has, for a good its action still serves. This brief sets out that rule, the natural-law reasoning behind it, and an architecture that could enforce it. It ends with a short list of open problems for researchers in philosophy, computer science, law, and safety engineering.