IEEE Technology and Society Magazine - March 2018 - 56

address Sharkey's technical arguments concerning the
shortcomings or possibilities of robots and computing
systems, I will critically examine his concerns and build
upon them from both a rule-utilitarian and a precautionary approach to argue that AWSs should be pre-emptively banned precisely because they are currently technically
incapable of complying with IHL and the discrimination
principle: if deployed now, they would not maximize the
reduction of harm to non-combatants, violating one of
the main tenets of just war theory and IHL.

Increased Attention on AWSs
It can be expected that research into AWSs will increase,
and that autonomous systems will be more ubiquitous
in military activities [3]. Many governments have prioritized the use of autonomous systems and military
authorities in countries such as the U.S. and The Netherlands have expressed favorable views towards autonomous weapons [4]. These weapons are less expensive
than manned systems, entail fewer apparent risks to
military personnel and could give nations a qualitative
edge. Another appealing notion is that autonomous
weapons could be programmed to comply with ethical
and legal norms [5], [6].
Over the last several years, legal, ethical and military
implications of AWSs have been repeatedly brought to the
attention of the international community by many
humanitarian organizations such as Human Right
Watch and the International Committee of the Red Cross
(ICRC). The issue was discussed in several meetings of
experts, most notably in those held within the framework
of the Convention on Certain Conventional Weapons
under the auspices of the UN Office at Geneva.
Advocacy and criticism against AWSs has also taken
ground; two important initiatives from civil society are
the Campaign to Stop Killer Robots,1 an international coalition that works to preemptively ban fully autonomous weapons, and a letter initially signed by 1500
prominent artificial intelligence (AI) and robotics re searchers calling for a ban on offensive autonomous
weapons [7].

Why Now?
Automatic weapons have been around for decades and
robotic weapon systems are anything but a novelty
as new teleoperated and telerobotic systems are constantly being developed and deployed in many areas
of conflict [8].
So, why is this type of weapon receiving specific attention now? One possible answer can be that because
up until recently even sophisticated weapon systems
retained a degree of human supervision over life-or-death
1

Campaign to Stop Killer Robots: http://www.stopkillerrobots.org/

56

decisions: drones are piloted from afar and have a high
degree of autonomous behavior, but there still is a
human person making the decision to fire a missile and
engage an enemy target. An automatic shotgun will
automatically load a new round and fire repeatedly for
as long as the shooter doesn't release the trigger. In both
these cases, a human intervention is necessary.
The current state of the art in AI and robotics has
reached a point where, according to many experts, it
is already technically feasible - or it will be feasible
within years - to deploy a killer robot that autonomously hunts and kills enemy combatants without
necessary human intervention. Precisely by making
human judgement unnecessary in the initiation of
lethal force, AWSs signal a paradigm shift in military
technology. The signatories of the aforementioned letter commented: "autonomous weapons have been
described as the third revolution in warfare, after gunpowder and nuclear arms."
Behind the opposition against killer robots, there is
the conviction that these autonomous systems "would
not be consistent with international humanitarian law
and would increase the risk of death or injury to civilians during armed conflict" [9]. If AWSs are not capable
of complying with international humanitarian law -
whose principles regarding the conduct in warfare are
rooted in Just War Theory - these weapons are both
illegal and ethically impermissible.

Between the Terminator and Reaper
The scope of this paper stretches between two extremes that mark what the discussion on AWSs is not
about. On the one extreme, there is the mythical Terminator, an anthropomorphic robot soldier equipped with
an intelligence that is near or above the human capacities. I will not be discussing this type of AI, also known
as strong AI (equal to a human's capabilities) or the
superintelligence (above human's capabilities). I will
stick to what is widely regarded in the robotic and AI
community as technically feasible in the coming years.
Perhaps, the current feasible limit on this extreme is
marked by what Bostrom [10] calls a "domain-specific
'superintelligence,'" i.e., one that is only smart within
one narrow domain. An existing example of a domainspecific intelligence is Google's AlphaGo, a computer
program developed to play the board game Go that in
March 2016 defeated Go master Lee Se-dol, one of the
greatest modern players. Autonomous cars could be
seen as another type of domain-specific intelligence,
albeit one that needs be much improved upon. In the
case of an AWS, the domain for this AI would be war, a
domain whose complexity is far greater than that of the
game Go or city traffic, itself already a very heterogeneous and complex domain.

IEEE Technology and Society Magazine

∕

MARCH 2018
http://www.stopkillerrobots.org/
Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - March 2018
