Slashdot Mirror
Mars Lander's Robot Arm Shuts Down To Save Itself
Cowards Anonymous passes along a PCWorld article that begins, "The robotic arm on the Mars Lander found itself in a tough position over the weekend. After receiving instructions for a movement that would have damaged its wrist, the robotic arm recognized the problem, tried to rectify it and then shut down before it could damage itself, according to Ray Arvidson, a co-investigator for the Mars Lander's robotic arm team and a professor at Washington University in St. Louis."
Wait, does this mean that the Mars Lander was programmed to comply with the Three Laws?
No. The second law translates to "Follow orders." The third law is "Don't get hurt (unless it conflicts with the second law)." If the lander had followed Asimov's laws, it would have followed the order and hurt its wrist.
In other words the Mars Lander performed as programmed. News at 11.
So the big question should be: Why are they sending it commands that could damage it? It's all good and well that it has some safty stops, but most machines do.
If you want news from today, you have to come back tomorrow.
I think it's amusing that after more than 30 years of Microsoft's quality control, when a computing device works as designed, it's a news worthy article. Think about it, I have a device that works as expected, can I be on the news too?
No, more like Matthew 18:8.
"If your hand or your foot causes you to stumble, cut it off and throw it from you; it is better for you to enter life crippled or lame, than to have two hands or two feet and be cast into the eternal fire."
Set your phasers on "funky"!
I've believed for a long time that laws 2 and 3 are the wrong way round.
You don't want an expensive robot to go breaking itself just because you're a bit careless giving it orders. Most devices are designed this way. Users are stupid. Even the smart ones. Even if I want to do something fairly harmless, like close an application without saving, the computer will stop me and check that's what I actually want to do.
what if this kind of code makes it into every piece of space equipment, and then by some fluke we are faced with the possibility of breaking a robotic wrist to deflect a space rock off an earth intercept course.
They should at least have a little clippy pop up and say "it looks like you want to break my robotic arm, are you sure you want to do that?" "are you absolutely sure?"
VLC FOR MAC IS DYING! IF YOU DEVELOP, PLEASE SAVE IT!!
The difference between the Mars lander and a car building robot is one of function.
The car building robot is programmed to do one task. It spends all day, every day, welding specific spots, on a car which is in a specific location.
The Mars landers have to content with an unknown environment, where they could be asked to do a wide variety of things, with any number of possible consequences.
Well, US robots rented its robots for a long time, I'm not sure they want them to break...
Yeah. Would you choose a neurosurgeon who pokes around people's brains in his spare time? I wouldn't.
You need to bound that forward-looking aspect of the law.
As the decision tree gets huge, just about any tiny action will eventually lead to tragedy, or [odius] being elected.
There is no real safety under the sun.
Get thee glass eyes, and, like a scurvy politician, seem to see things thou dost not.--King Lear
They're just lucky that the original system programmers, designers and testers that developed the fault detection code were better at their jobs than the mission programmers who fed the bad instructions to the lander. If it had been the other way around, misery and teeth-gnashing would have ensued.
We are the 198 proof..
...my Roomba, on a daily basis, recognizes stairs as a threat and refuses to fall down them. I guess I don't see the "big deal" here, sounds like a built-in protective measure worked as expected. The technology is no less awesome, but still, it functioned AS DESIGNED.
Unless you want the robot to sacrifice itself for you... Then order 2 preceding order 3 is VERY useful.
Such a case would be covered by the first law.
If you want to sacrifice the robot to save one of your other possessions, then the priority of the second law over the third is very useful.
Well, US robots rented its robots for a long time, I'm not sure they want them to break...
Leasing a robot was sufficiently expensive at the time that it more than covered for the specific repair. They also had the option to cease leasing to a particular client if he turns out to be destroying them on a regular basis.
In addition, a robot placing the orders of a human above its own self-preservation is a nice marketing point if you're trying to overcome the "Frankenstein Complex" that made humans afraid of them.
Warning: Opinions known to be heavily biased.