The Three Laws of Robotics: A joke on intellectual design? Or viable control method?

[BIos_Reaper]

I apologize dear furry science fiends! But I feel there just isn't enough discussion about Artificial Intelligence in this forum. I spend a lot of time thinking about sapience, intelligence levels, wisdom, free-will... etc. Comes with being in a psychology class. Free-will doesn't just "exist" it is learned, a human being that isn't taught doesn't turn into the average human we're used to interacting with, therefor I´d say humans have the capacitance for self-awareness but do not from the get-go, possess it, Tabula Rasa, if you will.

But back to point, as you all know, laws by nature are inflexible, therefor if circumstances change, or misinterpretations occur, the result of the machine minds view of the law changes considerably. Such as by protecting a human criminal by preventing certain documents from surfacing, because "a human will come to harm from this material." I find that this is an endless source of conversational material. I also think the anime series Armitage the Third or Armitage III introduces the conflict of humans vs humanoid entities, machines as living as you and I.

It is my personal belief that intelligence, that life, is not something human, indeed, not something at all biological in nature. "Life is not chemical, genetic or biological, to think so is simply wrong. Life is all that, which possesses the capacity, and ability, to self determinate, weather synthetic in origin or not."

Any discussions or opinions? I am intrigued to hear what you have to say, also, hi, I'm back! Well, somewhat, I haven't posted here in ages. Hope to hear from you soon!

[Nelle Kozera]

I prefer the "Obey at all costs" law, and leave it at that. Sure it could lead to some deaths among people but what ever, the dread robot uprising is avoided. Personally I think if your scared of machines getting out of control just hard code restrictions into there code via non rewritable storage mediums. granted it means the AI can't be transferred via the interwebs or other digital mediums but its meant for a physical platform anyway.

True life if you want to call it life is domain of the fleshy, synthetic life the domain for anything that didn't evolve via natural processes.

Also brownie points for mentioning one of the best animes i ever watched.

[Terrapun]

I'm not the best guy to talk about the “three laws”, as I have read very few of Asimov's works. But (in a true polish fashion, I will provide my input anyway) I always thought of them as flawed. They are more of a hurdle for a true intelligence than anything else, and you cannot build morality on laws and restrictions (Actually you could, but would it be a true morality or a mask ready to fall of at any time?). Education is far better in that regard as it gives a set of tools and ideas that can help even in some complicated situations. Not to mention the fact that AI operating under the “laws” is nothing more than a slave.

I, also, would not be so quick to judge what is a “true” life and what is not. Artificial or synthetic or what would you like to call it, life and, presumably, intelligence can come in the fleshy as well as metallic form, so where is the difference between the created life and the self organised one? In its origin? Is it that big of a deal? We are nothing more than an amalgam of molecular automata. Yes. We are the effect of natural evolution. And this makes us what? Better? Worse? In my opinion it's not enough of a reason for us to start differentiating life as “true” and “untrue”.

Late science fiction writer and, for the lack of better term, thinker, Stanisław Lem, had some interesting views on this subject. Many books written by him, explore the themes of machine intelligences, evolution and even beliefs. And how they can interact with one another and with other life forms. One of the quotes that are attributed to him is (I don't know if I got it right): “It doesn’t matter what you are thinking with, metal or custard, it's basically the same.” It sums his views about machine intelligence quite nicely. Although he seemed to think that true machine intelligence would be quite alien to us. Limiting the possibilities of interaction or making any form of interacting with, even understanding such a life form, impossible.

Well, I think that’s enough meddling for one day, pax.

[SliceOfDog]

Hah, I find this quite interesting, but the libertarian (if I'm not mistaken: Nelle) is arguing that there should be rigid laws built into robotics, and the socialist (or damn-well near enough) is about to argue that there shouldn't be. That is somewhat... curious.

In any case, I am afraid I disagree with both of you on several points here. I'll start with Nelle. The "Obey all laws" idea is, unfortunately, completely flawed. I actually wrote a short story about that exact idea, although it's pretty old so don't expect it to be amazing:

Short Story

The General peered hard into the jet black visor, but could only see his own reflection. He stepped back and viewed the men, if they could be so called, before him. They stood at attention, each and every one of them, unmoving, in five rows of twenty.
“Brilliant, aren’t they General?”
The General turned to look at the Captain, who stood beaming behind him.
“It took us years to get it right, but here they are; The lab’s finest”
‘The lab’s finest’ bore little resemblance to the human soldiers they were soon to replace. But for two arms, two legs and a head they could easily be machines; not an inch of flesh could be seen beneath their black armour, no facial features could be distinguished under their perfectly round helmets, they wore boots which would hamper the movement of any normal man in the field.
The General turned back to what the media had dubbed the ‘GM troops’.
“They’ll obey any order. Any!” the Captain went on “Can’t be beat in close-range combat if it comes to that, which you know it never will, if you’ve seen these guys shoot. Of course, America’s pining to snap ‘em up, but luckily this here’s the first batch, so we might just manage to ship ‘em out before the yanks. That’ll turn a few heads, eh, General?”
The General did not respond. Instead he pulled out a pistol, aimed at the nearest GM soldier and fired, all in one swift movement. The sound of the shot resounded off the walls, but the bullet had been a blank.
The Captain laughed nervously.
“Did you see that? Didn’t even flinch. What’dya reckon General? Shall I tell the guys in London ‘Job well done’?”
The General turned to face the Captain once again, solemn faced.
“I don’t like them” he said, curtly.
The Captain was visibly stunned.
“But... but General...”
The General pointed to one of the human soldiers, who stood near the exit.
“You there” he said, and though the soldier was positioned at the far end of the hall, the General’s commanding voice carried across the room with ease. The soldier stood to attention.
“Kill yourself” the General said, his voice bearing only sincerity. The soldier’s eyebrows rose, and he began to sweat.
“What... I mean... pardon sir? General?”
Without responding, the General pointed at the GM troops.
“Front row: Kill yourselves”
Before the Captain could protest, the black coated soldiers raised the barrels of their rifles to their necks, in eerie unison, and fired, slumping to the floor as one. The other rows didn’t so much as flinch.
“So which commanders would receive these new soldiers?” the General asked, calmly. The Captain stared at him, open mouthed.
“What... I... all... a... all of -”
“But they obey all other commanders?”
“What?”
“If they were assigned to an American commander, they must also obey English commanders, correct?”
“Yes... I suppose... yes”
The General nodded slowly.
“So if one commander should wish to take over the entirety of the western world, he need only ask?”
The Captain was silent.
“Tell London to begin work on something more useful” the General said “A talking dustbin, perhaps?”
He turned and walked away, his footsteps echoing loudly in the silence of the hall. The Captains simply stood, shocked, staring at the future of the military lying dead on the floor.

tl;dr version: Criminal tells robot to kill other people. Robot does so, because it must obey everything.
Someone tells a robot to disobey all later orders: what happens?
What happens with slang/mistranslations? If someone from England goes over to America and asks an American robot to "Help get me bladdered" what is the robot going to do? It has to obey, but is given what appears to be a nonsensical order. Does it do nothing? Does it attempt to 'guess' what the order means, so maybe rip out the guy's bladder, or get some bladders from a butchers shop and throw them at the guy?
What about words with dual meanings? "Robot, run me a bath". "Where would you like me to run it to, master?".
And that's really just scratching the surface. It's just riddled with flaws.

BIos, your arguments I disagree with on a more philosophical level.

Free-will doesn't just "exist" it is learned

I very much disagree. It is impossible to know whether free will exists (although I am inclined, for sanity's sake, to believe it does) but I certainly don't think that it can change. Either it does or it doesn't. We, as human beings, are either free to do everything, even though we might be heavily influenced to do one thing in particular (and indeed we're highly suggestible animals) or we're free to do nothing and every single choice is an illusion. Perhaps reality is pre-written, orchestrated by a higher power or something of that sort.
So I disagree that free will can be learned. Again, of the two I would like to think everyone does have free will, regardless of what happens or doesn't happen to them.

It is my personal belief that intelligence, that life, is not something human, indeed, not something at all biological in nature. "Life is not chemical, genetic or biological, to think so is simply wrong. Life is all that, which possesses the capacity, and ability, to self determinate, weather synthetic in origin or not."

This I do sort of agree with. I believe that AI could get to a stage where it could be considered 'alive'. However, the labelling is ultimately futile.
1) We are still arguing over whether animals have the capacity to 'self determinate', not to mention insects and concepts such as 'the soul'. Until we get that sorted out, there is no way that people as a whole will accept the concept of living machines. It just isn't going to happen.
2) Programmed machines are always going to be a direct result of their programming. Even if a machine was created to create other machines that create other machines ^50, and the final one creates AI that design other AI who build a third tier of AI, the whole process is still dictated by the programming of the first machine in the chain, which was at some point programmed by human or other biological beings. Perhaps the first machine is programmed to make all machines it creates pretend to be sentient, and also to program them to repeat that 'order' in their own creations. Therefore no action or reaction by a piece of AI can possibly be said, for definite, to be 'legitimate' and not a pre-programmed response.

The most important point in the above is that we can never know whether a machine is making a response or only making the appearance of a response (that is to say: can a machine be scared or only programmed to display what we associate with the biological emotion of fear?)

However, taking for granted that AI were to be created and could be proved to be legitimate, I would think that no 'laws' would be imposed upon them other than to follow normal human laws, and if anything they would just be reprogrammed if they disobeyed those laws.

[BIos_Reaper]

I am very happy with that response Slice! Your disagreements have made me think too. Free-will maybe isn't something you learn, but perhaps I should utilize a word like intelligence? For a human child left alone in the wilderness is still intelligent, all things can be intelligent, I mean a cat who figures out how to open a cabinet to get at the cat food early can be considered smart, and intelligence isn't something we can define as being purely human. Sentience is something we define as being human, or is it simply that we define ourselves a sentient because we create, communicate and advance?

I agree with you on the point Nelle made, that obey at all costs is a very bad idea and your story perfectly illustrates that. That is why the "Perfect" soldier will only ever be the soldier best able to perform his job without becoming blind to the nature of his superiors. Or to use the civilian term "All men in power must be doubted and questioned by their underlings to keep them in check." That is, absolute loyalty is a very bad thing to force onto anything. Maybe living machines will not appear just yet, but as Nelle told me, self-aware A.I.'s aren't practical, but since when have humans been solely focused on practicality? Have you seen some of the earliest aircraft? Outright silly is what I call them!

Also he says he doesn't want his computer to be able to disobey the orders he gives it, tell me why would a PC have an A.I.? I was talking beings that are seperate from commonday household appliances, why people make such inaccurate assumptions sometimes frazzles me. But also, you are correct, a robot must be able to understand a humans speech, and as such should probably have a dictionary system to catalogue and store words much like we humans do, that wont take much space at all. I think the oxford english dictionary is what... a few gigabytes? Maybe more... I do not know, it is a very thick and heavy book one can kill someone with XD.

Thanks for joining in also Terrapun! Your points have much merit to them :3 I for one think this topic is about to get interesting!

[Nelle Kozera]

Well what I said in conversation is still what I think Bios. But as for the topic at hand…

Now I care not personally one way or the other if AI’s are free or shackled, I was merely preposing a possibility for those that have issues about it.

Now on the concept of life I think I have a far more I guess rigid view on the definition of pure life and that it’s tightly confined to the biological or derivatives of life. The main reason being is what slice mentioned about them being alive. AI no matter their generation or complexity are still merely products of their programming and therefore just very clever simulations of life’s complexity and emotions.

Also he says he doesn't want his computer to be able to disobey the orders he gives it, tell me why would a PC have an A.I.? I was talking beings that are seperate from commonday household appliances, why people make such inaccurate assumptions sometimes frazzles me.

But as you said, you we never know the extent of an invention until the done. For me the logical extreme is the complete integration into the lowest levels of everyday life.

[13Swords]

Freefall addresses all of this. You guys should read it.