The U.S. military is working
very hard to develop autonomous robotic warriors. The U.S. is not
alone -- the technology is the wave of the future in warfare, and is being
pursued by dozens of countries worldwide. While expensive, weapons-toting
robots can provide deadly accuracy and protect a nation's human soldiers.
Already, the U.S. has extensively
utilized unmanned aerial vehicles in the war in Iraq, for both surveillance
and offensive strikes, with drones such as the Hunter
UAV shooting deadly missile strikes into enemy hideouts. The SWORD
robots, designed by the army and armed
with machine guns, patrol the streets in Iraq. Meanwhile a semi-autonomous
Bradley Fighting Vehicle, named the Black Knight, is being developed in the
U.S. Not to be left out, the Air Force states that it wants to have unmanned
heavy bombers by 2020.
The key feature among all these robotic killers that have been deployed or are
under development is that they need a human to pull the trigger. However,
there is growing sentiment among military circles that eventually robots should
be developed to be fully autonomous -- fighting and killing enemies, all
without human intervention. Exactly when and how such robots should be
legal and troubling moral issues raised were topics of
discussion at Royal United Service Institute's conference, "The Ethics
of Autonomous Military Systems", a summit of international scientists and
military officials held on February 27 in London.
The question "Can a robot commit a war crime?" was raised
during the conference. Such a concern -- that robots might
malfunction and target civilians or friendly soldiers -- remains a frightening
thought to many military men.
English Barrister and Engineer Chris Elliot explained carefully his thoughts on
the legality of autonomous robotic weapons systems in terms of international
criminal and civil laws. He points out that at a certain point the
robot's engineers can no longer be held reasonably culpable, and the blame for
errors resulting in catastrophic loss of life may come to rest on the shoulders
of the robot who committed the assault. He states, "We're getting
very close to the where the law may have to recognize that we can't always
identify an individual - perhaps an artificial system can be to blame."
The idea of a robot being charged with murder raises provocative questions
about punishment and the fairness of such measures. Elliot did not back
down from taking other hard stances on issues. He made it clear that
currently there was a clear legal burden for humans choosing to deploy systems
lacking in sufficient judgment. He stated, "Weapons intrinsically
incapable of distinguishing between civilian and military targets are
Elliot stated that robots should only be allowed to autonomously wage war when
they pass a "Military Turing Test." He explains, "That
means an autonomous system should be no worse than a human at taking decisions
[about valid targets]."
The original Turing test, developed by computer pioneer Alan Turing, states
that if a human is unable to tell in a conversation with a robot and a real
human, which is the man and which is the machine, then the robot conversing
with the human has achieved intelligence.
Elliot could not say when or how such a test could or would be administered,
stating simply, "Unless we reach that point, we are unable to [legally]
deploy autonomous systems. Legality is a major barrier."
Bill Boothby, of the UK Ministry of Defense, argued a slightly differing
perspective that if the situation was carefully controlled, an autonomous
fighting machine would not need to be quite as intelligent as Elliot's test
would require. Boothby hopes that by lowering the requirements, robots
could assist on the battlefield sooner as opposed to later. He describes
this stance, stating, "There is no technology today that can make that
kind of qualitative decision about legitimate targets ... [but] it may be
possible to take precautions in the sortie-planning phase that enable it to be
used within the law."
Boothby argued that in a way, human operators might be no better as they might
simply "rubber stamp" the robot's targeting decisions.
Boothby's comments were found more appealing to many military commanders who
wished for sooner deployment of such robots as the new SWORD autonomous
fighters or Foster-Miller's MAARS combat robot, which implements advanced technology to
reduce friendly-fire incidents.
The difference in opinions expressed between Elliot and Boothby are reflective
of the mixed feelings society holds about deploying independently operating
killing machines to warzones. There remain many fears in the mind
of the public, both realistic ones, based on practical assessment of the
current limitation, and fantastic ones, fueled by popular culture such as the Terminator
movies, which depict a future in which cold-blood lethal robots have turned
upon mankind. The issue is sure to only become more contentious with time
and technological advances.
quote: Now we have to enforce the Three Laws of Robotics
quote: Should we just throw away any thoughts on how they should work just because a science fiction writer thought it up first?
quote: In any case, I suspect we're decades away from any AI being smart enough to tell friend from foe.
quote: anyone who talks about starting a 'war' is immediately thrown into the closest mental health ward
quote: War will not have to be a 'reality of the future' if we stop allowing people to convince us that war is necessary, which in almost all cases it is not necessary.