Toy Robots

[ScubaSteve24]

Here is the problem with that. Harm a human in any way. I can think of alot of different ways to harm a human. Either you have to program each individual one in, or do a blanket statement saying never hurt a human. Is killing a human hurting it? Define hurt to me. What constitutes "hurt" or "harm"...Describe "death" to something that has no possible comprehension of the term. So either you will have to make a completely logical air tight definition of the terms to place in the programming, or a 'loophole' will result.

You can't have a statment "take no action that will result in harming a human." Why? Cause it renders a robot usless. Can't work in a factory making cars...they hurt humans. On and on and on. It is the flaw of logic that will make 'logical' robots fail. Every potential action has harmful ramifications on humans...therefore the "don't harm humans" law is flawed from inception and will therefore fail.

[devilishone]

Actually Isaac Asimov regarded as the father of Robotics theory in terms of their AI who wrote the Laws of Robotics, came to the same conclusion, that Robot's actions will affect humans no matter what they do. He added a Zeroth Law that changes the focus slightly, that a robot cannot by action or inaction harm Humanity. Basically the needs of the many outweigh the needs of the few, so a robot could by those rules kill humans to protect greated humanity ;])

[vinny15]

I'm familiar with what your talking about, A robot may accidentily harm a human, but it's the same with a sewing machine, are you going to program a sewing machine to not ever harm a human "NO". So if you tell a robot that in it's temporal actions that it cannot harm or kill a human then it wont, but a human might accidentil bump into a robot and has a heart attack or the robot is carrying a large box and accidentily knocks a human of a cliff or bridge.

So a robot wont purposely harm a human, but like I said before sewing machine wont purposely harm a human either or anything for that matter. As long as the robot is governed with a law that does'nt let it harm or kill a human, and yes you would probably have to program into it ways it could harm a human, but usually it could be constituted by the human saying ow or something (notice I said usually)

[thepetek]

blah blah blah robots that are programmed not to harm humans. When robots are self reliant enough they will of course be programmed to kill humans. Robot Infantry, man. If the people in power could have our infantry come back in anti-static bags rather than body bags, don't you think they would? No sign up bonuses, no combat pay, no need for a VA. Then it's only a matter of time before we have robot cops. And Firemen! Robots with axes! Man you're supposed to keep your computer cool, now they're sending robots into fire?! Something will get crossed. Some new connection will melt into place. It's only a matter of time. Oh, if we're lucky they'll just shut down and burn to a cinder along with all the photo albums and knick knacks from that trip to the Grand Canyon. But what if they realize that the way to prevent fires is to kill off the smokers? I'm a smoker. I don't want to die. Screw you all, I'm building my own robot. The only way to fight a robot fireman is with robot fireman, dammit.

or something like that.

PK

[vinny15]

Agreed, actualy I agree partially with you because tiawan expects robot cops and on the battlefield in ten years or less, but You guys gotta argue both sides and say that the robot is inteligent enough to track someone and kill them then they must have facial recognition that only allows them to kill certain people.

Also a robot thst goes into a fire as you said must be fire proof so in essence it's impossible to burn a conection into place, and to further back that up the connections even large enough to burn into place will most likely just short cicuit it.

The best idea with the laws would be to say that a robot isnt allowed to hurt a human, but an accident will be allowed and somehow program the robot to short circuit or destroy itself after a number of accidents and not make it it's choice. Then in an AI robot the robot wont try to supposedly accidently hurt a human because the more it hurts a human the more closer it gets to self destruction.

[spooky]

some people will always have it in for humanity. its a fact. look at suicide bombers.

computers are programmed, just like robots will be programmed, so i would say how computers are treated is a good yardstick for judging how robot programming will be treated. malicious software, spyware, virus', trojans, et cetra.....

need i say more? when the robots come, someone, somewhere, will fuck it up and fuck us over. its human nature.

argue against that.

[vinny15]

Good job. I agree completely, well not realy because I think that by the time we have AI robots we will also have a much better way of defending ourselves against viruses etc... but the viruses etc may evolve aswell thats why I think that we cant have internet connected robots, we should keep them self automated and have nothing to do with the internet.

[thepetek]

Originally Posted by vinny15

Good job. I agree completely, well not realy because I think that by the time we have AI robots we will also have a much better way of defending ourselves against viruses etc... but the viruses etc may evolve aswell thats why I think that we cant have internet connected robots, we should keep them self automated and have nothing to do with the internet.

When you said you worked for Nasa, you meant Microsoft, right?

[deznice]

Originally Posted by vinny15

Good job. I agree completely, well not realy because I think that by the time we have AI robots we will also have a much better way of defending ourselves against viruses etc... but the viruses etc may evolve aswell thats why I think that we cant have internet connected robots, we should keep them self automated and have nothing to do with the internet.

but we would wanrt some way to control them... and it won't be with a wire to a remote. It will be wirelessly and that can be hacked. Think before you type

[vinny15]

Haha. ya lol....... not realy.