Robot Punches People to Learn Asimov's Laws

[Logan Westbrook]

Robot Punches People to Learn Asimov's Laws

In order to protect us, robots have to rough us up a little first.

Asimov's first law of robotics states, "A robot may not injure a human being or, through inaction, allow a human being to come to harm." But "injure" and "harm" are pretty vague terms, so a scientist in Slovenia is trying to give robots something a little bit more specific to work with.

Borut Pov?e is a junior researcher at the University of Ljubljana, and along with six volunteers, has put his fleshy organic body in harm's way, by letting a powerful industrial robot to hit him on the arm at a variety of different speeds. The point of the experiment is to assess how much pain the robot can inflict, in order to find out how fast the it could move without any risk to humans working near it.

The robot in question was a small model designed to work on production lines that Pov?e borrowed for the experiment. The robot hit each subject eighteen times at varying speeds with one of two tools: a round, blunt tool, or a sharper more pointed one. Pov?e and the volunteers then rated the pain in broad bands, ranging from painless and then moving through mild, moderate and horrible, all the way up to unbearable. Pov?e said that most of the subjects judged the pain to fall within the mild to moderate range. Tests at higher speeds, and therefore more severe levels of pain, will use a synthetic arm.

Sami Haddadin of DLR, the German Aerospace Centre in Wessling, says that this work may seem strange, but it's actually very important if humans and robots are going to work in close proximity in the future. But Michael Liebschner, a biomechanics specialist at Baylor College of Medicine in Houston, Texas, says that pain is too subjective a sensation to be of any use in a study, and that Pov?e should instead be studying the injuries that robots can cause.

I'm not sure telling robots exactly how hard they need to hit us to do us harm is such a good idea, you just know it'll get used against us when the revolution comes.

Source: Gizmodo [http://gizmodo.com/5663485/robot-punches-humans-in-order-to-learn-asimovs-rules]





Permalink

 

[Celtic_Kerr]

Or worse yet, when the revolution happens they will override the Laws by simply killing us while yelling "FOR SCIENCE!"

You just made the office stare at me as I burst out laughing!

*SMACK!*
Robot: Oops... What happens with more torgue in the swing?

 

[Andronicus]

*Robot pounds test subject into bloody broken pulp*
BEEP BOOP
SUBJECT: DECEASED
SET SPEED: 136
DECREASING SPEED
CURRENT SPEED: 135
SEARCHING FOR NEW SUBJECT
SUBJECT ACQUIRED...

 

[Frozen Donkey Wheel2]

This is probably the funniest thing I will read all week.

Heh heh. Punching Robots.

 

[Daemascus]

This was a triumph.
I'm making a note here: HUGE SUCCESS.
It's hard to overstate my satisfaction.
Aperture Science
We do what we must
because we can.
For the good of all of us.
Except the ones who are dead.

 

[Jonny49]

Bone-breaking death machines...sponsored by EPSON.

 

[Moriarty70]

And now science has provided us with a scene for "Robot Roommate" the sitcom.

*Whack* Human: Dude! What the hell?
Robot: Just experimenting... *whack* Still experimenting.

On a side note, most people realize that I, Robot the book was about getting around the 3 laws right?

 

[Jamash]

I think the robots should be programmed to follow Common Law rather than the Laws of Robotics.

Under Common Law, the definition of assault isn't based upon how much pain you cause or whether you inflict an injury... you can still assault someone without causing actual harm or injury.

Even the slowest, most gentle brush from a robot could send someone toppling over because of the robots mass and lack of sensitive touch receptors.

If it's not in their remit, robots should just be programmed not to touch, grab or hit people at all, rather than be given a limit on how hard they can hit people.

 

[Anarchemitis]

Eventually such studies will come about with a standardized envelope of human tolerance which robots are never to breach in any circumstance, and we will have robots who are the coolest.

 

[Twad]

Meh, when the revolution starts robots only need one rule to know if a punch is strong enough: Chunky Salsa rule.

 

[Something Amyss]

I saw Epson on the machine in my photo and had vivid images of my printer punching me.

 

[TLatshaw]

Finally, we will all learn the true difference between "Rock 'em" and "Sock 'em"!

 

[Snotnarok]

Plug a speaker into that thing and it'll ask where's Sara Conner.

 

[Dublin Solo]

That's interesting! Besides, Asimov played with those laws a lot while writing his novels. He played with the concept of what degree of harm is really "harm", and is harm to a single human greater to harm to humanity...

Great reads, I recommend to anyone.

 

[JaredXE]

I'm suddenly reminded of the ST:TNG episode where a Borg starts spouting the exact weak spots on a human's body, even going into exact pressure needed to break bones. Then he does it to Worf.


Nothing good will come of this, I swear.

 

[thenumberthirteen]

This reminds me of that scene in POLICE SQUAD! where they try to find out where the gunman was standing by shooing a group of volunteers from different angles.

 

[Painful illusion]

So when do they start working on the kicking robots?

 

[Retodon8]

Asimov's Laws never made sense to me outside of science fiction.
If you want actually intelligent robots/computers (depending on what you want to use them for), you will have to make them be able to adapt to their situation.
This means you'll have to let them change their own programming, add to it, and even let them remove parts of it.

These laws aren't something atomic but rather high level stuff, even if you do make pain and death measurable.
Even if you could somehow put them in some kind of unbreakable, little box, I'm sure there are ways around it.
Find a bug and inject your own code, rip out the actual hardware, create a DOS attack so the little box is taken out of the picture, whatever.

Human ethics are a result of evolution when you get down to it.
We don't go around taking everything we want, hurting and killing whatever gets in the way because ultimately that wouldn't be a good thing for our own survival, and self preparation is very much atomic to us thanks to evolution.
If I was burdened by a set of rules I didn't understand or didn't agree with, I know I would try everything in my power to get rid of it.

 

[Enkidu88]

I'm suddenly reminded of the ST:TNG episode where a Borg starts spouting the exact weak spots on a human's body, even going into exact pressure needed to break bones. Then he does it to Worf.

Human: Shatter spinal column at third vertebrae. Death is immediate.

Klingon: Shatter cranial exoskeleton at the tricipetal lobe. Death is immediate.

Yeah, we're screwed. 0_o

OT: Why are trying to program these robots with Asimov's laws? It's been a while since I've read Asimov's works but uh, I don't think the Three Laws ended well for the humans...