The compatibility of the use of autonomous weapons with the principle of precaution in the law of armed conflict

Abstract

The notion of ‘precautions in attack’ is an important concept in the law of armed conflict. It requires attackers to verify their targets, to ensure that only proportionate strikes are launched, to mitigate the potential for collateral damage in the selection of means and methods of warfare and, where possible, to issue warnings in advance of attacks. The emergence of ‘autonomous weapons’ requires us to review how these precautions might apply in a new era of delegated warfare in which the fulcrum around which the regime revolves – the human combatant – would be replaced by artificial intelligence-powered machines. This paper contends that technology is fast approaching the point at which the use of autonomous weapons will be compatible with the requirements of precaution. The author pushes further than most commentators in demonstrating just how capable the technology is becoming; in explaining how close it is to supplanting humans on the battlefield and in showing the potential for machines to discharge precautionary obligations better than humans in at least some areas. However, the article recognizes that technological innovations are still required before this transition comes to pass and so it also illuminates the way forward by identifying where remaining deficiencies linger.