ALL SCI-FI

[Tom]

My issue comes with law #1

Law One - "A robot may not injure a human being or, through inaction, allow a human being to come to harm."

Just imagine that you were constrained by this rule. Lets also imagine said robot has enough AI to contemplate the rules. This gives the device an AI appx equal to our own intelligence. However, this same intelligence level gives the robot the ability to analyze scenarios a bit faster than us.

In my opinion, said robot would not be able to move due to conflicts.

Its based on the butterfly effect and said effect would incapacitate any protocol as strictly defined as rule number one.

The robot would be stagnated by the What If Scenario. Any action or movement by the robot would need to be 'approved' by its operating system before said movement could be initiated. How fast is it's processor?

The robot would analyze every possible outcome of its actions to the limits of its abilities. It could not 'take chances'. Chance might lead to a violation of rule #1. Therefore, before it would set to motion any action it would need to be assured that any coarse of action does not violate its prime directive.

Now lets imagine that it has found a coarse of action from its initial decision but time has passed from its initial calculation to its initial decision to act. No matter how much time has passed (for Super AI allow .001 second for trillions of calculations to be assessed) the robot now must reassess to account for life moving by during its processing. Sure life circumstances have not changed much in that time but they have changed. Now the robot must reassess every assessment it made to make sure nothing has changed that might allow a violation of rule #1. This process would repeat forever.

Law Two - "A robot must obey orders given to it by human beings except where such orders would conflict with the First Law."

The laws are already flawed with law #1.

Law #2 cannot work because humans are flawed. Humans do things without much thought to the consequences. A command for a robot to "go to the street and await further orders" would be simple for us but impossible for the robot. Not only due to the conflict of law #1 but because a device able to calculate rule #1 would also be able to calculate that humans have ulterior motives. So not only must it calculate law#1 for itself it must also attempt to justify the command to prove to itself that implementing Law #2 does not violate law #1.
Thus the time it requires to perform this calculation has changed the environment enough to require a new calculation to see if it is still safe to implement its action. It must also calculate to see if the travel to the street and its placement at the destination is not going to interfere events about to happen. "Going to the street and await further orders" might put it in the way of a child on a bicycle or some other out come.

Law Three - "A robot must protect its own existence, as long as such protection does not conflict with the First or Second Law."

The above endless computations and assessments would put its AI into a quantum loop of probabilities that might shut down processing and cause an overload. That overload would be harmful to the robots self interest. Laws 1 & 2 would break the robot from being able to function thus ending its existence.

Asimov later added the "Zeroth Law," above all the others - "A robot may not harm humanity, or, by inaction, allow humanity to come to harm."

Its very existence harms humanity. You see, I am part of humanity. You are as well. So is every other human in existence. If one of us is harmed, Humanity has been harmed. The fact that a robot even exists could cause a person to commit suicide then humanity has been harmed. If a group of people protest the robots creation and their passion against the creation has caused them to change their actions in any way that harms their ability to thrive then humanity has been harmed.

[bulldogtrekker]

You may be right about the 3 laws of robotics, but it is such a great device to write science-fiction stories with!



An equivalent idea is Arthur C. Clark's idea that geostationary satellites would be ideal telecommunications relays. A true fact today!

[Tom]

The world has already created robots with disregard to the Laws of Robotics

http://www.futurefirepower.com/swords-mili...re-weapon-robot

The SWORDS military robot is basically the first evolution of robotic weapons. What used to be mere science fiction in the 80???s has now become a reality, and 10 years from now, the SWORDS robot will be in the scrap pile. That???s because this tiny guy represents a proof-of-concept project to show the world, and the US leaders, that robotic tanks and such can and should be built. It goes to show that we can create offroad robots that can tote many different types of guns (the SWORDS can be outfitted with a rocket launcher, sniper rifle, machine gun, grenade launcher, and more), successfully locate the enemy, and take him out on command. Fortunately, the tiny robot isn???t autonomous. It is controlled by a human operator on a computer from far away, which is actually better than a fully autonomous platform, because if they controlled themselves???.Terminator anyone?







http://www.metalstorm.com/



http://www.youtube.com/watch?v=d8hlj4EbdsE


http://gizmodo.com/312443/robot-cannon-goes-berserk-kills-9

[Bud Brewster]

Mighty interesting post, Tom. Mr. Goostman says this --

[Tom]

It's not in the premise of the Laws that I have issue. It's in the wording.
A robot is not the same as an artificial entity.

No we can not build robots or AEs that could function with those restrictions. Will we be able to someday - perhaps. My question is Will we Want to?

I believe, and this is just my opinion, that when we can construct an artificial entity that is able to function the AI program will inherently be able to consider moralities.
A robot can not operate on 'intent'. An AI entity will.

Robots are preprogrammed and stupid. They follow their directives to their programmed outcome. If you insert an obstacle in front of it the robot will move around it, go thru it or stop.

In historical science fiction the term Robot describes an artificial entity. It's given human decision making skills to make it relative. Data from ST:TNG is not governed by the 3 Laws. That is what makes that concept so appealing.

I am not saying that the 3 Laws are not a great story device. I am saying they are not realistic to the nature of the beast. Historic Scifi using this device to explore the humanity aspects is wonderful. But now the Science in Science Fiction demands that we consider a better anology.

Thanx for the discussion!

[Pye-Rate]

I love the nonmilitary use of swords robots. From a semi take autoadjusting air suspension seat, add 3 sizes of electronic coolers [lunch, bait, fish], tackle box, rod reel gun rack, 5 miles from nearest road, FUN!

[Tom]

Hi Pye-Rate,
Interesting, Fun Ride eh?
Thanx for contributing to this discussion!

[Randy]