Autonomous Weapons Systems and the Future of War Written by Armin Krishnan This PDF is auto-generated for reference only. As such, it may contain some conversion errors and/or missing information. For all formal use please refer to the official version on the website, as linked below. Autonomous Weapons Systems and the Future of War https://www.e-ir.info/2013/05/27/autonomous-weapons-systems-and-the-future-of-war/ ARMIN KRISHNAN, MAY 27 2013 Military robots have been used in war in some primitive form since the beginning of the 20 th century. Only recently they have moved from a marginal role to the center stage of contemporary military and intelligence operations. The most technologically advanced armed forces have invested heavily into the development of unmanned vehicles and robots of all kinds, ranging from the now very familiar Predator drones to IED robots, heavy unmanned ground vehicles, unmanned maritime vehicles, automated sentry guns, micro-robots, malicious software bots for sophisticated cyber attacks, and even nano-bots. The common denominator for all these new types of weapons is that they are becoming more and more automated and ultimately autonomous. Mainly academics and peace activists have recently voiced strong concerns over the prospect of ‘killer robots’ roaming the earth and indiscriminately going after human prey. Although some of these concerns are legitimate, they are nevertheless strongly influenced by works of science fiction such as the Terminator or Matrix movies, which tend to reduce the complexities of these issues to the very simple formula that machines are evil and will ultimately bring about our own doom. It is argued here that at the core of the potential dangers of military robotics is the human factor rather than the intrinsic nature or the intrinsic limitations or even the capabilities of machines. The main concern should not be making our military robots behave ethically, but making sure that the human military decision-makers will have a strong sense of ethics and are constrained by an effective legal framework that prevents them from abusing the tremendous technological capabilities that will be at their finger-tips within a decade or so. Defensive vs. Offensive Autonomous Weapons Highly automated defensive weapons have been around since the early 1980s and they include highly automated air defense and missile defense systems, smart mines, and other area defense weapons. Although they can be considered ‘autonomous’ in the sense that they can independently identify targets, trigger themselves, and sometimes independently pursue these targets, they cannot be used offensively because they lack important capabilities such as mobility, cognition beyond an ability to identify a narrow set of targets, and endurance. A human has to switch them on, has to refuel and reload them, and has to do the troubleshooting. Defensive systems also tend to operate in environments that are typically devoid of civilians such as in closed off areas like international borders, the high seas, the deep sea, and outer space. In short, few people have considered automated defensive systems to be more immoral than other weapons used for similar purposes. In fact, greater autonomy in the sense of a greater capability for discerning targets can result in much better humanitarian outcomes. Offensive autonomous systems are much more problematic since they could operate with few limitations in geographic areas that are occupied by civilians and could accidentally cause war crimes by misidentifying targets. The problem is that offensive roles are very difficult for robots and remain at current time beyond the existing technological capabilities. However, in the long run it is foreseeable that autonomous robots could slowly move up the ladder from purely defensive roles to more offensive ones, or from a rudimentary capability of responding to an intrusion or attack to actively seeking out targets in an extended geographic area and attacking them preemptively. Thinking About Weapons System Autonomy Autonomy in an engineering sense refers to the capability of carrying out its core mission with little or no human supervision. Autonomy is a spectrum and is not easily definable. Depending on the complexity of the function or mission to be carried out by a robot, the machine would need to be more or less capable of understanding key E-International Relations ISSN 2053-8626 Page 1/3