Robot companions

Align

In the not too distant future, it's likely that there'll be humanoid robots well past the uncanny valley, capable of looking and acting realistically.
If you had one (or if that's too creepy to consider, if you encountered one at someone else's home or during a party etc), would you treat it like a person or like a thing?

lolfighter

Request for clarification: Would it BE a person or a thing? Are we talking <a href="http://en.wikipedia.org/wiki/Data_(Star_Trek)" target="_blank">Data</a> or <a href="http://en.wikipedia.org/wiki/EveR-1" target="_blank">EveR-1</a>? Or is that supposed to be unknown?

Align

That is the point of the thing, not to know it!
references aside, it'd be a robot, and so technically a thing, programmed with set rules to react in certain ways in certain situations. But this might be hard to believe when it's crying its eyes out after you called it as much.

lolfighter

Well, if it follows set rules to react in certain ways in certain situations without the ability to deviate from them in ways its creator couldn't anticipate, then it isn't a person. It's a thing.

But I assume that the spirit of it is that I don't know whether that is the case or not. The safe bet, then, is to treat it like a person. I am confident that if it is merely a thing, it will give itself away sooner or later. And if it is a thing, treating it like a person does no harm, whereas treating a person like a thing would be offensive.

Of course, treating like any OTHER person is out of the question. There'd always be a difference. I wouldn't ordinarily question a human person about the nature of their being, whereas this might be the first thing on my mind if I encountered a potential AI.


Here's a different take on the subject: What if it wasn't a robot, but merely a voice and microphone?

locallyunscene

Definately treat it like a person. I mean, as lolf said, that's the safest way. It doesn't really matter what it looks like as long as it seems intelligent enough. For all you know that toaster oven talking to you has a fleshy brain inside, assuming we're in a place technologically where we're creating near Turing-test capable machines.

Rob

There are some who don't think the Turing test is all that useful. Specifically, there's more to intelligence than carrying on a conversation. Intelligence is kind of like obscenity: you know it when you see it. In a sense, we as people don't act in any way that is outside of our programming, either. I'm of the opinion that a sufficiently complex robot would be indistinguishable from a human, at least from an intellectual perspective, if given the challenge of evolving like we did.

As far as how you would treat a robot, well, it would depend on whether or not the robot has emotions. If it developed emotions for some reason through evolution, like we did, then you'd need to keep that in mind when you interact with it, least you make it mad or something. If it never needed emotion and never developed it, it probably wouldn't matter how you treated it. It would matter more how it evaluated you in regard itself: are you helpful or dangerous.

aNytiMe

You don't treat robots like humans, that would destroy the need for men to find a real woman and would destroy humanity. Unless androids had real, fully functional female genitalia. <img src="style_emoticons/<#EMO_DIR#>/nerd-fix.gif" style="vertical-align:middle" emoid="::nerdy::" border="0" alt="nerd-fix.gif" />

Also, having sex androids will totally decrease competition for men with game and it'll basically be a threesome at hello.

TychoCelchuuu

I'd treat it like a human because otherwise I would feel awkward and self-conscious. That's my practical answer. As for the philosophical answer, which is harder, I'd say that it would be a good idea to err on the side of caution and basically assume that we've created a new life form. If we made half cats, half dogs, we wouldn't treat them badly just because they're artificial. Same deal with Robo Dude.

locallyunscene

How timely: <a href="http://www.guardian.co.uk/technology/2008/oct/05/artificialintelligenceai" target="_blank">article</a>{guardian.co.uk} on the same subject.

It poses an interesting question, once it gets to the level we're talking about, do we have the right to switch it off?

BlackHawk

well that sets it back to the, is it a THING or a PERSON, in which case conflicting morals would be involved, But would you let it get that far?

lolfighter

What do you mean, let it get that far? Can you elaborate?

t3hd3r3k

imho, its just a set or programed ai. IT would not be a person and would be incable of feelings. So as i am saying, i would not treat it as a person but as a object. Also, someone had to throw in the other side of this argument. (forgive spelling)

lolfighter

If we're gonna play devil's advocate, let's play devil's advocate: If I can't tell the different between this thing's behaviour and yours, how do I know YOU are a person and capable of feelings?

Align

I think I made a doodoo.

The original idea I had was, if you <i><b>knew</b></i> it was just an AI, would you treat it like a human (because it would feel weird to treat something that looks and acts like a person like it wasn't) or like an object (because that would be technically correct)? Which is pretty much opposite to what I asked. I forget...

lolfighter

Hm. Yes, it IS necessary to clarify that an AI is an object and not a person if you wish that assumption to be the basis for the discussion, since otherwise this is still very much a subject of debate.

X_Stickman

I would treat any robot capable of holding a conversation as a human.

If only because one of the most disturbing scenes in any show/film I've ever seen is in the Animatrix "The Second Renaissance" where they beat that womanbot to death with a sledgehammer.

Man even thinking about it disturbs me.


But yeah. All human interaction is based learned responses to situations. Granted they are *amazingly complex*, but an advanced robot would be a simplified version of us (or a sufficiently advanced robot would be exactly like us).

Thansal

<!--quoteo(post=1690231:date=Oct 14 2008, 11:22 AM:name=Align)--><div class='quotetop'>QUOTE(Align @ Oct 14 2008, 11:22 AM) <a href="index.php?act=findpost&pid=1690231"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I think I made a doodoo.

The original idea I had was, if you <i><b>knew</b></i> it was just an AI, would you treat it like a human (because it would feel weird to treat something that looks and acts like a person like it wasn't) or like an object (because that would be technically correct)? Which is pretty much opposite to what I asked. I forget...<!--QuoteEnd--></div><!--QuoteEEnd-->

See, that still leaves this question:
If it is significantly advanced, so what? If it has equal/near/greater reasoning, learning, and emotive capability then a human, then so what if it is an AI?

If we encounter life from some where out there (Solar system/galaxy/universe/whatever), that has all those things (And is friendly), do you decide that it isn't human and thus treat it like an object?

Now how does an alien differ from a true AI?

Rob

<!--quoteo(post=1690354:date=Oct 15 2008, 05:54 AM:name=Thansal)--><div class='quotetop'>QUOTE(Thansal @ Oct 15 2008, 05:54 AM) <a href="index.php?act=findpost&pid=1690354"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->See, that still leaves this question:
If it is significantly advanced, so what? If it has equal/near/greater reasoning, learning, and emotive capability then a human, then so what if it is an AI?

If we encounter life from some where out there (Solar system/galaxy/universe/whatever), that has all those things (And is friendly), do you decide that it isn't human and thus treat it like an object?

Now how does an alien differ from a true AI?<!--QuoteEnd--></div><!--QuoteEEnd-->

It doesn't, but that wouldn't stop us from treating an alien as a thing, too. The question could be whether or not the AI is sufficiently 'like us' to merit treating it 'like us.' Eliminating all the things you said we're the same on, the AI still wasn't born to a human mother. A similar debate then might be in vitro humans who also are somehow carried to term outside of a mother. But then, the AI isn't fully flesh and blood, either.

I would imagine that, after something like this actually happens, there will be a consensus on ethics for it. Anyone who would have to work with an AI would need to go through an AI Ethics class. AIE 101.

lolfighter

It's another case of "similar but different." We wouldn't (and shouldn't) treat AIs and sapient aliens like humans, because they're not. But that doesn't warrant treating them like objects. An object has no rights, except by extension of its owner. The same couldn't possibly be applied to sapient aliens, or to AIs.

Rob

Another aspect that might change the situation is the number or organization of the AI(s). Are we talking just one, or the first one, where the mechanic's creation suddenly says "I hurt" here? Or are we talking a whole community of AIs, if they indeed organize themselves that way? This gets really complicated, because it's easy for us to assume that intelligence can only exist in the stand-alone single unit version that we are.

So, ants in an anthill, individually, they're just following orders from a hive mind or some such other nomenclature. Taken as a whole, they can engage in mass coordinated warfare. Is the ant collective intelligent? Does it have feelings?

If you say no, then consider this: Each of the neurons in your brain is very dumb and simple. Voltage comes in, if it's above a threshold, the neuron fires and sends voltages to any number of other neurons. There's no apparent communication except for this simple mechanism between neurons. Are you intelligent?

Intelligence may depend more on the complex interactions of a system than any individual's processing power. Considering that, you may act differently toward a single AI that simply performs tasks while learning opposed an entire group of the same AI acting together. If the AI group encounters an unpredictable problem (read: outside the design scope of the programmer), comes to a consensus solution, and implements that solution successfully, it may be hard to deny that the AI, or at least the AI community, is intelligent.

The motivation to treat intelligence with respect is at least partly based in self-interest: intelligence is unpredictable, uncontrollable, and therefore dangerous.

Prefix

<a href="http://news.bbc.co.uk/1/hi/technology/7624425.stm" target="_blank">http://news.bbc.co.uk/1/hi/technology/7624425.stm</a>