On May 15, ICRAC’s Heather Roff delivered the following statement on operational issues to the informal “Meeting of Experts“, gathered to discuss questions related to “lethal autonomous weapons systems” from May 13 to May 16 at the United Nations in Geneva, Switzerland.
Operational statement by the International Committee for Robot Arms Control
Convention on Conventional Weapons Meeting of Experts on lethal autonomous weapons systems
United Nations Geneva
15 May 2014
Thank you Chairman.
Distinguished delegates, on behalf of International Committee for Robot Arms Control (ICRAC), I would like to raise three potential operational points concerning the future deployment of lethal autonomous weapons.
First , INTEROPERABILITY:
The issue of interoperability is crucial. Military systems must be interoperable and secure to be effective. To deconflict a battlespace and minimize the likelihood of friendly fire, these systems must be able to recognize and communicate with each other. However, this will be increasingly difficult because of limitations in bandwidth, the probability of hacking and cyber attack, and the inability to properly field test large-scale, interoperable, multi-domain combat to ensure reliability and predictability. Connectivity and communication is often degraded in warfare, and it is likely will only intensify in all domains populated by autonomous weapons systems.
This will only be exacerbated in allied or Joint Operations. As many states have discussed here already, the requirements of secrecy for the successful protection of their systems will undermine the ability for allies to engage in joint operations. Their autonomous weapons systems must be able to “talk to” one another, and if this is the case they must either have similar operating systems or states must share this information.
Second, UNOCCUPIED SPACES:
There are no unoccupied spaces. There is only less probability of encountering civilians or civilian objects, and the assumption that one can keep any domain separate from another is pernicious. Indeed, it suggests a kind of authoritarian population control. The accidental downing of Iranian flight 655 shows that the Air/Sea distinction is not a dichotomous divide. Naval systems may fire upon aircraft, and may, as in the case of the Aegis system misidentify a civilian airliner as an incoming fighter jet. The history of mine warfare should also warn us of the difficulties of keeping civilians out of supposedly solely military space like a minefield. The trend of contemporary combat is not one of neatly separated spaces, but the urbanization of hostilities and complex mixing of civilians and combatants.
Furthermore, even if limited to a low probability domain, such as submarine warfare, such systems may be incapable of upholding the Laws of the Sea, in particular duties to rescue or prohibitions against unnecessary force. Moreover, given that communication is difficult in underwater domains, the increasing likelihood that these machines will have no communication with command and control is operationally dangerous. Precursor systems under development include underwater submarine mines designed to loiter and patrol for months; it would be difficult to assign command responsibility for the “launching” of these systems.
We are pleased to see the focus on costs. The assumption behind the creation of autonomous systems is that they will be cost effective. However, the reality of waging war requires large uses of initial force to gain dominance over one’s adversary, particularly gaining air dominance to protect all other assets. A state may use autonomous systems to gain tactical air dominance, but this will prove prohibitively expensive and capital intensive. It would require such complex systems far and above the costs of the F-35. If however states do not utilize autonomous air systems, and opt for current manned combat aircraft, then we have a question of why there is a need create the systems in the first place.