Accountability Gap, Autonomous Weapon Systems and Modes of Responsibility in International Law

53 Pages Posted: 31 Mar 2016

See all articles by Thompson Chengeta

Thompson Chengeta

Harvard Law School; Midlands State University, Faculty of Law

Date Written: September 30, 2015

Abstract

In most circumstances AWS are incapable of complying with rules of International Humanitarian Law and International Human Rights Law leading to violations of important rights like the right to life. The question that follows is who is responsible for the violations of an Autonomous Weapon System. In this paper, I focus on the challenges of accountability that are posed by AWS and the possible solutions to such.

AWS without ‘Meaningful Human Control’ are unpredictable on the battlefield or wherever they are used. In the event of them violating the law – violations that are not intended by the person deploying them – it is not clear who is legally responsible, thereby creating an accountability gap. Accountability is important in international law because where there is an accountability gap; the victims’ right to a legal remedy is adversely affected. There are four forms of accountability that I am going to discuss in this paper: individual, command, corporate and state responsibility. Under individual and corporate responsibility, there is civil and criminal liability.

In summary, the arguments I make in this paper are: the above mentioned forms of accountability are complementary to each other; they are not alternatives to the exclusion of the other. For example, if AWS create an accountability gap as far as the individual criminal responsibility of those deploying AWS on the battlefield is concerned, that specific gap is neither closed by suing the responsible individuals under civil responsibility nor holding the manufacturing company liable under corporate responsibility.

Under individual responsibility, as long as there remains the possibility of AWS acting in an unpredictable manner, they may present an unresolvable challenge as far as the establishment of the accused person’s mens rea is concerned. I also argue that the proposed system of ‘split-responsibility’ over use of a weapon – where responsibility is divided or shared between the fighter and other persons involved in the production of AWS like manufacturers – is not only foreign to international weapons law as the lex specialis on the use of weapons but also inappropriate and hence unwelcome.

As for command responsibility, I argue that it is inapplicable to the relationship between AWS and those deploying them. No analogy may be drawn between the relationship of human commander versus a human subordinate and that of the human fighter versus a robot. The continued referral of a person deploying AWS as a commander gives a misleading impression that AWS are somewhat combatants or fighters. AWS must be developed in a manner that they remain weapons in the hands of a fighter who is liable on the basis of individual responsibility in cases where crimes are committed. It should not, and must not be a case of a commander and subordinate where the notion of command responsibility is invoked. Command responsibility is only applicable to the extent of the responsibilities of a human commander over his or her human subordinates involved in the deployment or use of AWS.

Persons involved in the production of AWS have their own responsibilities in the designing, manufacturing, selling and transferring stages. This is where corporate responsibility also comes into play. I note, however that although corporate responsibility is a sound form of accountability, it has an inherent weakness of putting the onus on victims to bring cases against robot corporations which in some cases are registered in foreign countries thereby presenting insurmountable difficulties for the victims. Victims will not only face monetary challenges in terms of legal costs but will also be confronted by jurisdictional challenges.

State responsibility is like an umbrella to all the forms of responsibility mentioned above; covering and enforcing corporate responsibility at the design stage of AWS up to selling or transferring stage; enforcing individual and command responsibility when the weapon is finally used on the battlefield or law enforcement situations. As one commentator has observed, when considering accountability over the actions of AWS, state responsibility ‘is the frame of reference for considering other forms of international responsibility’. From a state responsibility perspective, I also acknowledge the genuine fear that AWS may make it possible for some states to deploy force against other states in non-attributable ways.

In conclusion, I recommend that the only way to address the accountability challenges that are presented by AWS is to make sure that humans exercise ‘Meaningful Human Control’ over weapons. Where ‘Meaningful Human Control’ is exercised, AWS will remain mere weapons in the hands of the warriors – that is exactly what they should be. In short, however, I propose that the notion of ‘Meaningful Human Control’ over the use of a weapon is only satisfied where the control that a fighter exercises over a weapon is to such a degree that the actions of an Autonomous Weapon System are entirely his – the system depends on the control of the human fighter to execute the ‘critical functions’ like the decision as to who to kill and legal calculations on the lawfulness of an attack.

Keywords: Autonomous Weapon systems, accountability gap, killer robots, responsibility

Suggested Citation

Chengeta, Thompson, Accountability Gap, Autonomous Weapon Systems and Modes of Responsibility in International Law (September 30, 2015). Available at SSRN: https://ssrn.com/abstract=2755211 or http://dx.doi.org/10.2139/ssrn.2755211

Thompson Chengeta (Contact Author)

Harvard Law School ( email )

Harvard Law School
1541 Massachusetts Avenue
Cambridge, MA 02138
United States

Midlands State University, Faculty of Law ( email )

P Bag 9055, Senga, Gweru
Gweru, Midlands +263
Zimbabwe

HOME PAGE: http://www.msu.ac.zw

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
373
Abstract Views
1,506
Rank
147,540
PlumX Metrics