By Michael Binnington
After the last informal meeting of experts in Geneva on killer robots (or as they prefer to call them “lethal autonomous weapon systems”) wrapped up it is an appropriate time to take stock of what we learned from the conference. A lot of ground was covered in Geneva, too much to cover in one short blog post, but there were a few ideas that received a lot of attention that are worth mentioning here.
First and foremost the idea of ‘meaningful human control’ got a lot of attention from all sides in the debate. So what is meaningful human control and how does that impact the debate on killer robots? Simply put, meaningful human control means that a human will always be the one that makes the decision whether or not to use force. There are three ways in which these systems are often described: human ‘in the loop’, human ‘on the loop’ and human ‘out of the loop’. A system with humans ‘out of the loop’ is the type of system that can target and use force without any human control and is the type of system that the Campaign to Stop Killer Robots seeks to ban. Systems with humans ‘on the loop’ give humans the ability to monitor the activity of the weapon and stop it if necessary. However, these systems may not furnish the decision maker with enough time to assess the information reported by the weapon. Finally, systems with humans ‘in the loop’ are more akin to traditional weapon systems, where the decision to use force rests firmly with a human operator.
The discussion of meaningful human control was linked to discussions about whether or not it was ethical or moral to delegate life and death decisions to machines. Some criticize this approach on the basis that meaningful human control isn’t a legal standard, or is too vague, but that criticism misses the point. This moral and ethical consideration is at the heart of the debate on killer robots; if only strict legal standards were applied then the ability and function of the technology would begin to determine how it is used. Strictly applying legal standards may approve the use of killer robots in areas that seemingly have no impact on civilians such as in outer space. Once such a precedent was set it would be difficult to stop the full use of killer robots.
After meaningful human control, the arguments made against a pre-emptive ban on killer robots formed a consistent theme throughout the conference, no matter the specific subject at hand. The refrain goes something like this, “We don’t know how this technology will evolve, so a pre-emptive ban could deprive the world of potentially useful technologies”. There is a concrete example of this not happening (the ban on blinding laser weapons), and various other treaties with dual-use implications have proven that banning a class of weapon does not adversely impact commercial or industrial activity. The Chemical Weapons Convention, which was discussed, provides a good example of how an export-control regime and competent verification can stop the spread of chemical weapons, while maintaining the ability of states to develop chemical industries.
Clearly then, neither of these two things should stop us from a pre-emptive ban on killer robots. As a co-founder of the Campaign to Stop Killer Robots, Mines Action Canada encourages all of you to engage with the issue and to advocate for a ban with your friends, family, local politician and anyone else who wants to listen. An easy way to start would be signing and sharing our petition to Keep Killer Robots Fiction here: http://killerrobots-minesactioncanada.nationbuilder.com/.
Michael Binnington is a M.A. Candidate at Norman Paterson School of International Affairs and a Research Associate at Mines Action Canada.