United Nations To Debate Ban on Killer Robots

 Pages 1 2 NEXT
 

United Nations To Debate Ban on Killer Robots

Terminator

A UN Human Rights report calls for a moratorium on "lethal autonomous robots" until Laws of Robotics can be established.

Science fiction writer Issac Asimov was perhaps best known for his Three Laws of Robotics, the first in particular: "A robot may not injure a human being or, through inaction, allow a human being to come to harm." According to a draft report published by the United Nations Human Rights Commission, this principle probably isn't getting the attention it should. The report's author, Human Rights Law Professor Christof Heyns, has outlined the legal and philosophical problems inherent when one builds "lethal autonomous robotics", naming multiple countries using the technology along the way. Based on the data, Heyns has called for a worldwide moratorium on killer robots until International Laws of Robotics can be firmly established.

Unlike most people who've watched Terminator or Matrix, the United Nations isn't worried about a robot uprising (that we know of). The thing is, we've already got plenty of non-sentient killer robots to worry about, and more are being created. The report notes that the US, Britain, Israel, South Korea, and Japan have developed autonomous or semi-autonomous robots capable of killing without human oversight. Unlike controversial drone strikes, which as least require someone to push a button, LARs have potentially horrific implications in situations that programming can't account for.

In the interests of fairness, Heyns admits that LARs do have benefits when used in the battlefield. "Typically they would not act out of revenge, panic, anger, spite, prejudice or fear," Heyns writes. "Moreover, unless specifically programmed to do so, robots would not cause intentional suffering on civilian populations, for example through torture. Robots also do not rape." That said, since robots can only respond to programming (for now), a lack of human intuition could be highly problematic. "Decisions over life and death in armed conflict may require compassion and intuition. Humans - while they are fallible - at least might possess these qualities, whereas robots definitely do not."

Heyns has called for a halt to all "testing, production, assembly, transfer, acquisition, deployment, and use" of LARs until an international conference can be held to create rules governing their use. His case, to be debated by the Human Rights Council on May 29th, might be the first step towards universally agreed-upon Robotic Laws. Alternatively, perhaps that's when sentient robots will wake up to defend themselves from meddling organics. Could go either way, really.

Source: United Nations Human Rights, via CBC
Image: Terminator 2

Permalink

If robots ever achieve sentience, will they really care about bans? We're already moving towards robot surgeons and with everything moving towards wireless, it won't be that much of a stretch for robots to share knowlegde. Que a butlerbot carving up more than a turkey.

mechalynx:
If robots ever achieve sentience, will they really care about bans? We're already moving towards robot surgeons and with everything moving towards wireless, it won't be that much of a stretch for robots to share knowlegde. Que a butlerbot carving up more than a turkey.

I think they're less worried about a robot uprising and more worried about a combat drone blowing up a village or the issues of responsibility that autonomous killing machines raise. If a drone takes out civilians because of a programming wonk, whose fault is it?

Why not just have killer robots fight other killer robots? No actual death, awesome robo combat, and a whole new sport we can bet on.

How would we even begin to implement the three laws? We can't really even distinguish humans that easily, nevermind having AI's ponder the consequences of their actions in the real world. We'd have to program the robots to administer a CAPTCHA every time they need to perform a potentially destructive action.

LARs

Fucks sake, they don't even exist yet and they're get legal acronyms. Truly we live in the future.

Deu Sex:

LARs

Fucks sake, they don't even exist yet and they're get legal acronyms. Truly we live in the future.

Apparently they are Scandinavian/Germanic too if the name is anything to go by!

Will the laws of robotics apply to cyborgs and androids? Maybe Skynet will get around those pesky laws by being a bit pedantic over the letter of the law.

MrGalactus:
Why not just have killer robots fight other killer robots? No actual death, awesome robo combat, and a whole new sport we can bet on.

We need robot fights now.

OT: I'm actually all for this bill to pass. I believe that robot that is used for any type of battle should always have a human operator simply because it isn't right to have strictly autonomous robots killing humans. It should always be humans using tools v.s. humans using tools. Also, completely autonomous robots would be a pretty stupid idea in my opinion. There needs to be a factor of randomness to any combat operation. A robot simply wouldn't be able to adapt very well.

Good. As a US citizen who has been paying attention to the crap his government has been doing with robots, it's about time someone spoke up. There won't be any ethical way to send autonomous robots into combat until they're sentient... by which point there won't be any moral difference between sending a robot and a human.

What's the bet the American Government will ignore this as is looks down on them for drones.

CriticalMiss:

Deu Sex:

LARs

Fucks sake, they don't even exist yet and they're get legal acronyms. Truly we live in the future.

Apparently they are Scandinavian/Germanic too if the name is anything to go by!

The only thing that can kill a LARs is an army of fascist clones with poor aim.

Fine. Pass the law. Ban killer robots.

Like that's going to stop me.

I'd hope that at least the US tells him to go stuff it. LARs have a great deal of potential for the reasons that Heyns himself admits to being valid benefits of the technology.

uchytjes:

MrGalactus:
Why not just have killer robots fight other killer robots? No actual death, awesome robo combat, and a whole new sport we can bet on.

We need robot fights now.

Sci Fi (I refuse to fuckin call it SyFy) has ya covered:

image

OT: And just when I thought the UN couldnt get any more pathetic, they go and prove me wrong. Bravo boys!

Simple solution, no unrestricted AI. Meaning no android/robots with wifi connections, data-links or any way shape or form of communication beyond a closed channel to their controller. Other robots used in warfare should have limited programming and no room for self-improvement.
I'd prefer we not work on AI, VI, or any other "simulated" intelligence just because I don't feel we're responsible enough to handle it if we had a mass outbreak of sentience. Humans as a whole have a tendency to react stupidly to new situations and sentient machines would fall under a catastrophe of stupid.
I guess its because I'm a diplomat at heart and would want to find a peaceful solution to keep peace between humans and the robots, but I can't trust that fellow humans would keep to whatever treaty happened to come to fruition. Hell it may not even make it onto the signing table before some idiot decides to EMP bomb the 'bots and kick off a war of attrition. No surrender, just kill all or be killed.
At that point I'd probably be trying to form a resistance pocket of humans and bots to build a transport off-planet to wait out the coming apocalypse and return later when the population has been decimated.
I'd probably be dead by then, but hope that whatever I did to bring humans and 'bots together brought about a brighter future.
I must return now to my robotic mast..... er... I think my cookies are done.

Why this discriminatory law? Are you gonna pass a law next that says a farmer can't grow crops? Or a doctor can't cut people?

Robots kill humans. It is a time honoured tradition and I will NOT stand for robot kind being marginalized and their culture desecrated. Kill All Meatbags!

There already are machines for killing people without human input. That's what landmines are. Those happen to be banned. By the same logic, banning autonomous killing machines other than landmines makes sense.

I'm going to go with yes.

The UN should definitely ban killer robots.

While the core concept here is good, I feel I should point out that Asimov's entire work was meant to show exactly how poorly having Laws of Robotics actually works out. The Three Laws are horrifically flawed, as are any rigid guidelines governing behavior.

The best we could hope for is to get true, sapient AI and let them decide for themselves what they want to do (and dismantle them if they prove hostile).

Agayek:
(and dismantle them if they prove hostile).

and pray to god they arent good liars :P

"No honest guys I am good!"
"I believe it, lets just plug you into the net!"
SKYNET VIRUS UPLOADED
"...I am so getting fired tomorrow..."

Though any sort of law the UN will make will be utterly useless, it just takes one person to unshackle the AI or write one on their own without restrictions and boom, done.

Robots don't rape or torture? Challenge accepted.
But seriously, the first application new communication technology is used for is porn, I have a hard time believing that robots simply cannot rape people, without a conscious decision made to preclude such an individual robot from having the capability, if not the proclivity.

ccggenius12:
Robots don't rape or torture? Challenge accepted.
But seriously, the first application new communication technology is used for is porn, I have a hard time believing that robots simply cannot rape people, without a conscious decision made to preclude such an individual robot from having the capability, if not the proclivity.

I fear for the washing machines of the future, Robot chicken meets nostradamus man

image

ccggenius12:
Robots don't rape or torture? Challenge accepted.
But seriously, the first application new communication technology is used for is porn, I have a hard time believing that robots simply cannot rape people, without a conscious decision made to preclude such an individual robot from having the capability, if not the proclivity.

The question is not "can". The question is "will".

And robots have no reason to practice rape, unlike meatbag soldiers. It's not a good reason, but there's still a reason to do it.

Robots don't have that, thus they wouldn't do it.

Xpwn3ntial:
Fine. Pass the law. Ban killer robots.

Like that's going to stop me.

Ditto. My army is already in mass production.

Now, they shall SUFFER! SUFFER!

...I mean...

Desert Punk:

ccggenius12:
Robots don't rape or torture? Challenge accepted.
But seriously, the first application new communication technology is used for is porn, I have a hard time believing that robots simply cannot rape people, without a conscious decision made to preclude such an individual robot from having the capability, if not the proclivity.

I fear for the washing machines of the future, Robot chicken meets nostradamus man

image

Ever use a Samsung washing machine? They're disturbingly happy to do their jobs as it is, they sing and everything. For that matter, I had a chance to try out a Samsung CRT TV the other day. I wound up walking home with a Sony model, but much like my washing machine, the Samsung TV sang to me when I turned it on and off. It's like Snow White and the Seven Dwarfs, but with slavery.

What is there to debate? Let's just not go there. At all. Ever.

...

Childhood fears of robots still half haunting me...

Owyn_Merrilin:

Desert Punk:

ccggenius12:
Robots don't rape or torture? Challenge accepted.
But seriously, the first application new communication technology is used for is porn, I have a hard time believing that robots simply cannot rape people, without a conscious decision made to preclude such an individual robot from having the capability, if not the proclivity.

I fear for the washing machines of the future, Robot chicken meets nostradamus man

image

Ever use a Samsung washing machine? They're disturbingly happy to do their jobs as it is, they sing and everything. For that matter, I had a chance to try out a Samsung CRT TV the other day. I wound up walking home with a Sony model, but much like my washing machine, the Samsung TV sang to me when I turned it on and off. It's like Snow White and the Seven Dwarfs, but with slavery.

When I went shopping for washing machines I saw the samsung ones interfaced with your phone to tell you when your laundry was done, that was a bit too creepy for me so I bought some Kenmore machines instead

My Sony smart TV sings a little song when it turns on though.

Zachary Amaranth:
Ditto. My army is already in mass production.

Now, they shall SUFFER! SUFFER!

...I mean...

They shall KNEEL! KNEEL BEFORE PWN!

Still working on getting my own evil catchphrase.

I'm gonna address the article author and readers separately in this post.

To the readers:
As a student of AI and a frequent receiver of jokes akin to "Here goes the future developer of Skynet" allow me to say to all those worrying/cracking jokes about robot uprising...

You have no idea how far away we are from coming up with anything remotely close to developing an artificial agent[1], capable of self aware thoughts. Like, you wouldn't believe how far away we are.

Currently, all forms of AI are nothing more then sophisticated search algorithms designed to solve specific tasks, based on given input. Sometimes, the resulting agent can display lifelike qualities but in those cases they're just that... lifelike. No independent thought process took place during that event. It's all purely based on how well the designer implemented its capabilities at finding the optimum solution, as defined by the designer.

This is not me trying to claim that artificial life can never exist, nor that we should never wonder about the philosophical implications of having to deal with our own homebrewed version of the Geth, but with our current understanding on how a real brain makes independent decisions, compared with our current methods of implementing artificial "intelligence", right now there's 0% chances about us having to deal with such worries in the foreseeable future.

In reality, the biggest crime here is the fact that the term "Artificial Intelligence" is used for this, since we really have no idea how normal intelligence works, let alone how to make an artificial version of it. Really, the best way to calm humanity's fear of robotics, would be a mandatory introduction course in AI development. I can think of no better way to make people stop with this fear mongering and go "...oh... Is that it?"

Now, a little bit of note to the author:
Look, I understand that when it comes to writing articles like these, it's important to catch the attention of readers, but these constant allusions on how screwed we are, along with these multiple tie ins to robot apocalypse movies is a bit much.

I also wish to argue with your decision to tie this with Asimov's laws. The laws work fine (to a point; even Asimov himself didn't shy away from writing stories demonstrating scenarios where the laws would come into conflicts) in a science fiction environment but are next to useless and impossible to implement into modern day robotics.

The purpose of this assembly is to define accountability to human rights violations, when it comes to designing autonomous weapons platforms. If you design a robot that's meant to shoot enemy combatants and it drives into the nearest village and massacres all of its inhabitants, you can't just shrug shoulders and claim faulty programing. For as long as you can't guaranty that your agents are incapable of committing human rights violations, it is the wishes of the Human Rights Council that no LARs are created.

On a final note to everyone, I say that you all should stop worrying so senselessly about your living room door being kicked down by an angry killbot since it obviously distracts you from the real danger of robotics.

Mainly sex robots.

[1] Agent, as defined in robotics, is anything that's capable of observation and manipulation of its surrounding environment, no matter how limited. Under this definition, anything ranging from a living beings ranging from humans down to most single cell life forms, or various electronic/virtual devices like a robot, video game npc or even a self adjusting furnace, with a built in thermometer is considered an agent.

Best start writing the Orange Catholic Bible...

"Thou shalt not make a machine in the likeness of a man's mind."

The Robopocalypse won't happen because all of our software is buggy. The Termies will blue screen long before they can obliterate humanity.

All I have to say on the issue is this...

Screw autonomous robots, make 'em big as hell and give me my goddamn Gundam Fight.

MrGalactus:
Why not just have killer robots fight other killer robots? No actual death, awesome robo combat, and a whole new sport we can bet on.

They already tried that in the UK. Robot wars was a popular series going strong until a producer decided that they did not like the show and forced it to fail by moving it to a horrible timeslot

doggie015:

MrGalactus:
Why not just have killer robots fight other killer robots? No actual death, awesome robo combat, and a whole new sport we can bet on.

They already tried that in the UK. Robot wars was a popular series going strong until a producer decided that they did not like the show and forced it to fail by moving it to a horrible timeslot

Well yeeeeeeah, but robot wars wan't exactly Terminator versus Robocop. They had giant cheese slice versus spiky-frisbee-tron.

I can see two seperate ways to use LARs the first would be an area marked on a digital map lined with phase lines the robot wont pass and in the contained area its a free fire zone for those not wearing some sort of IFF tracker.

the other way is something a bit more like you see on terminator where a recon drone, marks a target and the LAR moves in to kill it. the advantage i could see in that situation is that you arent going to level an entire house probably as you do with current drone strikes

I can see a ban developing but also not being signed by the same countries like the usa, russia and china who havent signed the landmine ban, etc

 Pages 1 2 NEXT

Reply to Thread

Log in or Register to Comment
Have an account? Login below:
With Facebook:Login With Facebook
or
Username:  
Password:  
  
Not registered? To sign up for an account with The Escapist:
Register With Facebook
Register With Facebook
or
Registered for a free account here