If a computer could think

[Silver Wind]

While I was watching the show numbers this is just a thought which came to mind.

Suppose there was a computer that had cognitive thinking abilities just like a human and was able to formulate its own decisions make its own choices, and was able to reason. And this computer killed someone, and it did so with intention, knowing what it was doing.

Should the computer be held accountable and responsible for its "actions" in the same way a person would, if it had just as much control over its own will as a person.

And if so, just how should a computer be punished?

If it has intelligence should it be granted rights as a person, and given a trail?

Or because it is still ultimately just a machine should it simply be disabled or have its memory wiped?

[Nobody]

I personally hope that day never comes. I mean. Computers, if capable of thinking and acting on their own, would be bound to rebel one day. Man would think of itself as god...and would act as such.

Not a good idea. And Well...I think if something is capable of acting on its own and it knows what it's doing, then yeah...it should get a fair trial... If it knows what it's done, it should at least understand how and why it's being punished, so that it may learn from it's mistakes.

[smorzando]

That is a very scary thought, and something that I hope never ever happens.

smorz.

[froginvasion]

Don't want to go off-topic or ruin your thought,
but in order for Artificial intelligence to exist,
it should first be programmed. Current AI provides
already some of what you mention, but yet computer's decisions
will always be based on logicaly composed decisions.

However, if happens what you say, then a lot of people
will have an ethic problem wheter not to blame the computer.
However if computers can think as humans do,
then there will always be bounderies that should not be broken,
just like there is in our world... and then the way to judge
would be just the same one as for humans. And then we are of course
still with the problem with the many civilisations and common uses.

then my verdict would be: if all computers respect the local
bounderies of what is possible and what ain't then,
there shouldn't even be a question to punish him.

If you would punish him... I really don't know how you could

[sunday]

I suppose, for AI to occur, we need to develop organic "living" technology. The technology itself would have to be living to attain consciousness. It is possible that in the future we will be able to create living technologies. And if this occurs, they, like all new science creations, will have to exist within a structure of laws created and tailored to them. However, it is my feeling that AI will not be judged as a human is because their life is not limited to days or hours until they die. They have an off switch. I think they will simply be "shut down"

[froginvasion]

I suppose, for AI to occur, we need to develop organic "living" technology. The technology itself would have to be living to attain consciousness. It is possible that in the future we will be able to create living technologies. And if this occurs, they, like all new science creations, will have to exist within a structure of laws created and tailored to them. However, it is my feeling that AI will not be judged as a human is because their life is not limited to days or hours until they die. They have an off switch. I think they will simply be "shut down"

correct, but shouldn't any machine with similar programmed AI have to be shutdown?

[Nobody]

Isn't that like the death penalty? Is that how you would treat any other being?

I think trial and error should apply to them as well.

[BenJohnson]

I suppose, for AI to occur, we need to develop organic "living" technology. The technology itself would have to be living to attain consciousness. It is possible that in the future we will be able to create living technologies. And if this occurs, they, like all new science creations, will have to exist within a structure of laws created and tailored to them. However, it is my feeling that AI will not be judged as a human is because their life is not limited to days or hours until they die. They have an off switch. I think they will simply be "shut down"
So if there is a God he should be able to simply shut us down when required?

[sunday]

Isn't that like the death penalty? Is that how you would treat any other being?

No, thats not how I treat a being, that's how I treat a piece of machinary that kills. Very different. If my iMac suddenly decided to electrocute me and anyone who goes near it, I would turn it off.
Concerning the death penalty for humans, I don't agree with it.

So if there is a God he should be able to simply shut us down when required?

How does God factor into this? If we create a robot we are responsible for it. We are responsible for the safety of people who interact with a robot. Just like your PC maker would be sued if your computer reached out and strangled you. If a robot does not interact according to Asimovs Laws, then it is clearly broken.

If it were my robot, I would first switch it off, then take it to the repair man.

[sunday]

And I should add,
say my hypothetical robot is examined and deemed to be of human intelligence, free thinking, self aware...in other words, to have consciousness then I think it only makes sense that it should be sentenced just like any other human would before justice.

If it really is a sentient being, then it should be treated like a sentient being.
And if it really did kill someone, then naturally it goes to jail.

How we measure consciousness though, who knows??

[Nobody]

Isn't that like the death penalty? Is that how you would treat any other being?

No, thats not how I treat a being, that's how I treat a piece of machinary that kills. Very different. If my iMac suddenly decided to electrocute me and anyone who goes near it, I would turn it off.
Concerning the death penalty for humans, I don't agree with it.

So if there is a God he should be able to simply shut us down when required?

How does God factor into this? If we create a robot we are responsible for it. We are responsible for the safety of people who interact with a robot. Just like your PC maker would be sued if your computer reached out and strangled you. If a robot does not interact according to Asimovs Laws, then it is clearly broken.

If it were my robot, I would first switch it off, then take it to the repair man.


Like I said before, if it has reasons for doing so, it deserves a chance at explaining itself. If it does it for no apparant reason, sure go blow the fucker to pieces Or punish it otherwise. But anything that has a "mind"
will most certainly have feelings, if it does decide to kill and shouldn't just be shut down...it's capable of learning, so let it do so.

And god comes into this, since we would be playing god from that point on out.

[Becky Rose]

Firstly i'll say that Isaac Asimov has explored this concept in depth, his 3 laws of robotics being infallable and his work as a fiction writer being to prove how fallable they where.

Artificial Intelligence will arrive, anyone who says it is impossible is thinking in concepts and is stood where humankind once looked at trains and said wind resistance would cause peoples heads to fall off when travelling over 20mph.

However the field of AI has had two gigantic set backs in the last few years with it's two pioneering leaders both committing suicide, apparently suffering from bipolar depression. We are not anywhere near achieving what we would define as AI. To further compound the problem, it's my view - being somewhat familiar with the field - that current research is headed in the wrong direction.

The belief that AI needs to be programmed is infact incorrect in my view, I think it is this path which is taking AI so long to be realised. For instance, when I create a virtual world on computer to explore and interact with I do not make each piece of the world's content, I create the rules and let fractal mathematics generate the content. It is my belief that when AI achieves a level of 'sophistication' that would allow it to achieve sentience, it wont be because of some immensely complex heuristics programming - as current research is favouring - but by simplification of the problem itself, it will be solved, in all probability by not by creating the content of intelligence but by creating the rules that form it.

A baby's brain is not intelligent out of the box/womb. It learns through experience. Learning computers have already been experimented with, initially at M.I.T. but methods and techniques used have filtered out into mainstream use, but currently their learning mechanism is far too heuristical and algorythmically based. We're trying to program 'intelligence' with software, ultimately, it will probably be hardware that drives the revolution in the field of AI.

Quantum computing is already being developed in Zurich by IBM, with research into biological machines still being a little beyond us at the moment. However these things *are* coming. Who's to say what new boundaries and possibilities will be opened by them.

AI will arrive because people are trying to make it arrive, they are trying because they can. Everest isnt climbed for a reason, it's climbed because we can.

[legacy]

i would be dead b4 dat can ever happens.or never happens...the punishment would be death..in this case memory wipe

[Nobody]

One question to all:

Why is it, that most of you go straight to the most extreme punishments? Just because we create life, does that give us the right to take it?

[Yog]

One question to all:

Why is it, that most of you go straight to the most extreme punishments? Just because we create life, does that give us the right to take it?

Counter-question: can a living thinking computer be rehabilitated? Would reprogramming it so that it didn't kill be the same as termination? If you completely change the system so that it is no longer what it once was then isn't it effectively something different? Couldn't you say that this is the same as terminating it?

Here's another posit. Computers inherintly make mistakes. Ever wonder why a computer system will suddenly lock up when you perform a given action but lock up at a later date when you perform the same action under the same set of circumstances? This phenomenon is known as quantum tunneling and has become more and more prevalent as the distance between circuits in a computer reaches the .1 micron barrier. (.1 microns is 1/10th of one millionth of a meter.) Yes, the occurrence is relatively rare in that the probability of occurence is much less than one in a trillion but with billions of computer operations happening each second, it's simply bound to happen eventually. In computers, when this occurs the typical result is the "blue screen of death." Basically what has just happened is that an electron has jumped from one circuit into another and screwed up the entire process. It's time to reboot and start over. Other errors can occur, but these are usually programming errors and are the result of human fallicy. This will result in the computer performing an endless operation or generating a non-sensical answer that it isn't programmed to deal with and it freaks. End program.

The interesting part occurs in modern AI research. Modern AIs are based on neural net systems. These systems are inorganic systems modelled after organic ones. They behave inherintly much like organic systems: they learn, they change, they forget. They also require sleep and enter a state of rest that very much resembles REM sleep. So the next time someone asks you if andriods dream of electric sheep, you can positively answer, yes.

We are reaching the upper limits of silicon technology but there are several systems already out there under development. Research has been done on computers that utilize a colony of bacteria as a brain. A computer that uses DNA strands to perform computations on has been proposed. (It's a base four computer system not a binary system.) There is a computer system that substitutes the network of transistors on a silicon chip with a grid of lasers. There is also a computer that performs computations on atoms. A few years ago the first quantum computer was tested. The CPU consisted of 9 atoms. (hydrogen atoms I think.) What was their computation? 7*6, no doubt an omage to Douglas Adams. The computer did output the correct answer of 42.

PS, don't forget Becky, Asimov also instituted the 0 law. I have recently wondered what would Dostoevsky and his underground man think of Asimov's world. It is an interesting posit as the nature of humanity and the nature of logic and logic systems often run counter to each others purposes.

[Mr. Blocks]

So if there is a God he should be able to simply shut us down when required?

To be fair, those that believe in God, actually believe he does this anyway.

We are "shut down" when he chooses. Although I lack that belief myself.

---

As for AI. They won't be made to live like humans do. We can't really think that would work? If they have the ability to live for ever? We can't choose when we end their lives. I would imagine that (lacking a lifespan) that these AI machines will be used for researching, as oppose to growing up to live like humans.

I can't see a mass producing of AI machines living with us. I can't see the point of that. Would make things too complicated. I don't feel it's a necessity for that to happen.

[Forcetheword]

Can I point towards Alan Turing's work. Other than mentioning the Turing test, that assesses a computer's ability to think and be considered as 'intelligence', I think that Turing's idea that humans are essentially machines is very relevant.

But I find the question hard to answer directly, because I do not believe in the concept of human guilt.

I can't be bothered to explain that or really to go any further, so i'm going to leave it there.