backtop


Print E-mail del.icio.us 64 comment(s) - last by wordsworm.. on Apr 1 at 10:27 PM


Would it be immoral to deploy future autonomous versions of war robots, such as the human-controlled US Military's SWORD robot, shown here?  (Source: Gizmondo)
Campaigners says deploying autonomous murderous robots is not a smart idea

A group that has long focused its lobbying efforts on stopping the proliferation of land mines is turning its attention to a surprising new target:  war robots.  In the first known instance of a non-government group protesting against war robot technology, the London-based charity, Landmine Action, hopes to ban autonomous killing robots in all 150 countries currently bound by the current land mine treaty.

While all machine gun-packing robots currently are human-controlled, the U.S. Department of Defense has expressed interest in deploying autonomous robot warriors onto the battlefield in the near future.  Last month DailyTech reported that Noel Sharkey, a robot researcher at Sheffield University, expressed controversial concerns about the ethics of autonomous war robots.  He stated that such robots might be capable of "war crimes".

Sharkey's speech inspired Landmine Action to take action against the war robots.  Richard Moyes, Landmine Action's director of policy and research, says the fight against autonomous killers is not a policy switch.  He says the organization has already fought cluster bombs, which use infrared sensors and artificial intelligence to decide when to detonate.  Landmine Action believes that taking the targeting decision out of human hands, and putting it in a machine's is a deadly one.

Moyes explains, "That decision to detonate is still in the hands of an electronic sensor rather than a person.  Our concern is that humans, not sensors, should make targeting decisions. So similarly, we don't want to move towards robots that make decisions about combatants and noncombatants."

The organization hopes to sway the International Committee of the Red Cross and Amnesty International, two leading organizations in war ethics lobbying.  Landmine Action is spurred on by Sharkey's comments, including his statement that, "We should not use autonomous armed robots unless they can discriminate between combatants and noncombatants. And that will be never."

Many in the robotics community express agreement with Sharkey's sentiment.  Peter Kahn, researcher on social robots from the University of Washington, states that he believe Sharkey to be correct and hopes that robotics researchers will stop taking government money to design war robots.  He argued to his colleagues at a conference on Human-Robot Interaction in Amsterdam, "We can say no.  And if enough of us say it we can ensure robots are used for peaceful purposes."

However, most in the robotics community feel this is impossible as most robotics research is funded by the Defense Department.  Says one anonymous U.S. researcher at the conference, "If I don't work for the DoD, I don't work."

Some robotics researchers disagree with Sharkey and feel that robots could make the perfect warrior.  Ronald Arkin, a robot researcher at Georgia Tech, says that robots could make even more ethical soldiers than humans.  Arkin is working with Department of Defense to program ethics into the next generation of battle robots, including the Geneva Convention rules.  He says that robots will react more ethically as they have no desire for self-preservation, no emotions, and no fear of disobeying their commanders’ orders in case of bad orders.

Many, however, remain skeptical of the wisdom of deploying increasingly intelligent robots onto warzones across the world.  They point to the many science fiction scenarios, which depict humanity at war with killer robots of their own creation.  While this may seem farfetched, the issue of war robots is becoming a serious one that the world's brightest minds are trying to grapple with.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

Quite simple
By FITCamaro on 3/31/2008 10:17:33 AM , Rating: 5
quote:
We should not use autonomous armed robots unless they can discriminate between combatants and noncombatants. And that will be never.


Uh its actually quite simple. If you're shooting at the robot, you're a combatant. If you're not, you're not a combatant.




RE: Quite simple
By FITCamaro on 3/31/2008 10:20:23 AM , Rating: 5
Besides, before an autonomous fighting robot was ever deployed, it would go through extensive testing and simulated situations. And the good thing about a robot is it doesn't know the difference between simulated and real. You give it a training weapon (paintball gun or laser gun) and tell it to act like it should in the field. The results say the rest.


RE: Quite simple
By BMFPitt on 3/31/2008 10:27:16 AM , Rating: 5
But what if it gets struck by lighting and starts thinking on its own?


RE: Quite simple
By monitorjbl on 3/31/2008 10:29:53 AM , Rating: 3
INPUT!


RE: Quite simple
By threepac3 on 3/31/2008 11:02:25 AM , Rating: 3
Short Circuit!!


RE: Quite simple
By HOOfan 1 on 3/31/2008 12:41:24 PM , Rating: 2
Numba 5 ALIVE


RE: Quite simple
By therealnickdanger on 3/31/2008 2:57:06 PM , Rating: 3
Wouldn't you like to be a pepper too?


RE: Quite simple
By AnnihilatorX on 3/31/2008 2:09:20 PM , Rating: 2
What about a software bug?

Who's fault would that be then?
The programmer?


RE: Quite simple
By GaryJohnson on 3/31/2008 6:16:55 PM , Rating: 2
What happens in other industries when hardware and software problems have deadly consequences?

For example, in the Ford/Firestone tire tread seperation fiasco, who ended up being at fault?


RE: Quite simple
By Nik00117 on 3/31/2008 11:29:00 AM , Rating: 1
I agree 100%. You forgot when a human does training several things he knows for a fact
1. Hes not in any real serious danger, he knows 100% that he will not die, at least thats typically the case.
2. He knows also roughly on what to expect.

A robot is different, your taking out the human nature aspect of it. Which is something which needs to be desired for. Want to ask why? Well lets say your on a partol in Iraq your convoy gets blown up and their was a girl in the hummer that got hit and shes now dead. You two sercetly had a love relationship type deal going on, your the commander of the sqaurdan so you order the surronding houses to be cleared. Your marines get pumped up run and go on a mass killing spare.

Robots won't have that issue, their buddy robot gets blown up they will mark it down as 1.2 million dollar loss and keep fighting.

HOwever your asking for another stack of problems, its a software/hardware combotion. Granted programed and developed by some of the best of the best, however far from perfect. One could ask, what if something programmally, or hardware like went worng. What if one robot went on a killing spree due to a software fault? Sure its unlikely, and sure we would of thought of it. But I who has planned many network upgrades, computer installations, upgradegrad across mutliple machines, development of programs etc knows for a fact their typically is that one variable which was never quite addressed. This was typically created by two actions or more which were designe dto prevent or do something else. IT relaly is a complicated science, on this scale that is. So to assume that this system will be fault free is fool heartly at best insane at worst. And many may argue "Well if out of 50,000 missions it misses up once thats ok" well ok, of course the American public won't here about 49,999 missions. So needless to say it'd be considered a failure that one mistake.

Do I believe its a good idea to have robots fighting our wars?

YES! I do, if we could just have "robot" wars that cost a lot of money i'd be happy. At least humans aren't being killed.

Do I believe robots which act upont heir own accord is a good idea?

YES!, but wait! We aren't anywhere close to the development of that. Lets wait a couple decades before we truly consider it.


RE: Quite simple
By MrBlastman on 3/31/2008 11:54:48 AM , Rating: 2
We are decades away for sure before a true AI is realized with any high-degree of safety.

We can though, currently employ robotic units that are remotely controlled by humans. I think that is where we are at now and they show great potential in this area.

They might not completely replace manpower as of yet due to SA (situational awareness) constraints, but they can dramatically reduce risk-exposure to human life in tasks that could create a high-probability of cessasion of existance.


RE: Quite simple
By theapparition on 3/31/2008 12:03:37 PM , Rating: 2
quote:
if we could just have "robot" wars that cost a lot of money i'd be happy. At least humans aren't being killed.

And who are the robots killing? Other robots? What happens went the losing side runs out of robots? Do they surrender or have to fight the robots? When does it end? Not so clear cut, is it?

Don't get me wrong, I absolutely am in favor of battle robots, but the point of war is to be "messy". So messy that it is undesireable, and such that both sides would rather compromise than to go down the "messy" path.


RE: Quite simple
By MrBlastman on 3/31/2008 12:11:00 PM , Rating: 2
I always thought the point of war was to swiftly dispatch your enemy with as little loss to your own side as possible?

Brute force to yield ultimate societal control. If you slaughter all your enemy (army and citizens), what do you have left to rule over? Burnt out cities and demolished technology? Perhaps resources if the region provides them but I'd want more to have a "persuaded" populace to do my bidding without weakening my own.

Messy or not, if the Geneva Convention did not disallow it, I am sure assassination or political subterfuge through outside pawn placement (think governmental control asserted through direct influence or elusive implantation of candidates) would be sought after more.

Absolute destruction creates unneccessary chaos.


RE: Quite simple
By FITCamaro on 3/31/2008 12:35:09 PM , Rating: 2
Assassination needs to make a comeback. A sniper can achieve what it takes months to do with a ground war. They sure don't have any problems sniping us.


RE: Quite simple
By snownpaint on 3/31/2008 2:33:27 PM , Rating: 2
Black Bag operations.
The wars without war.


RE: Quite simple
By Master Kenobi (blog) on 3/31/2008 2:36:44 PM , Rating: 2
Western world likes to put "rules" in place that have no place in a war. War is by definition without rules.


RE: Quite simple
By Christopher1 on 3/31/08, Rating: -1
RE: Quite simple
By MrBlastman on 3/31/2008 4:51:40 PM , Rating: 3
I think your statement applies to Islamists in the middle east equally.

They have been raising quite a stink lately about their beliefs in many countries throughout the world.

European nations are changing overnight due to these happenings.

So, before you go and point your finger at America being the ultimate evil due to our influential nature... consider first that there are others outside of us that would like to do the same as well.

I'm not saying we're innocent (we aren't, we've really screwed up actually), but...

Wars are a neccessary evil with our current genetic makeup. Until violent tendencies have been completely wiped from our chromosomes, man will continue to conduct violence on themselves.

It is sad, but it is reality.

Once the reality is accepted, you can make a choice:

a. To either do nothing and face repercussions.

b. To take action via violent, massively destructive actions.

c. To take action through swift, decisive means that target purely the evil itself with minimal collateral damage and minimal involvement.

d. To attempt to solve said course with reason, negotiation and logic.

Unfortunately d. is unobtainable by the majority of the world for any significant duration of time due to genetic difficulties, so we end up resorting to a, b or c.