New Weapons, Proven Precedent: Elements of and Models for a Treaty on Killer Robots
While some states have suggested that the cutting-edge nature of fully autonomous weapons will significantly complicate the treaty process, drafters of an instrument on the topic can look to existing international law and principles for guidance. These weapons systems present distinctive challenges, and no single source constitutes a model response, but creating new law from scratch could unnecessarily slow the progress of negotiations.
This report provides precedent for each of the treaty elements and shows that constructing a legally binding instrument does not require an entirely new approach. Earlier law and principles, often driven by similar concerns and objectives, can inform the structure of a treaty on fully autonomous weapons, and when negotiations start, facilitate crafting of language. The existence of relevant models should make it legally, politically, and practically easier to develop a new treaty.
This publication can be found on the Human Rights Watch website here.
International law and non-legally binding principles of artificial intelligence (AI) provide ample precedent for the elements of a new treaty. Lessons from the past can and should be adapted to this emerging technology.