Toy Dog Robot "Doesn't Want" to Shut Off?

[INRM]

I remember hearing about some Japanese robot-dog that when they asked it to shut off, it allegedly "didn't want to" and wouldn't shut off.

Does this mean it's sentient and is aware it's alive and doesn't want to be turned off?

[Jimbo07]

Even if true, did somebody just write a routine to disable a shutoff?

[MRC_Hans]

The barrier to sentient computers (whether in robots or otherwise) is that all useful computers are still von Neumann machines or variations thereof. In other words, they are completely algorithmic. They will ultimately only react the way they are programmed to react (even if that algoritm includes some virtually unpredictable randomization routine).

Of course, one may philosophically claim that humans (which we define as sentient) are also ultimately algorithmic.

Hans

[Ripley Twenty-Nine]

Sometimes when I click on 'Shut Down' in Windows, it doesn't respond.

Should I assume that Windows is sentient and doesn't want to be shut down, or that it's a problem with Windows?

[Magic 9-Ball]

Do you mean sentience or AI?

It could mean it has AI, and was looking for a fire hydrant for a little electrical discharge. When ya gotta go, ya gotta go.

[Safe-Keeper]

As Hans already said, computers don't do things they've not been programmed to do. If the dog refused to shut down, it's because it's been programmed that way.

Or, of course, that it was misunderstanding input, which also happens, bugs being what they are.

[Marquis de Carabas]

The Most Beautiful Machine (http://www.kugelbahn.ch/sesam_e.htm) doesn't want to be turned on.

[Safe-Keeper]

Awesomeness:D. That was hypnotizing to watch.

[tsg]

I remember hearing about some Japanese robot-dog that when they asked it to shut off, it allegedly "didn't want to" and wouldn't shut off.

Does this mean it's sentient and is aware it's alive and doesn't want to be turned off?

Furbies have a habit of occasionally telling you "No' when you ask them to tell you a story or sing a song. It's part of the programming.

[Graham Jackman]

Does anyone else remember the text based game based on "The Hitchhiker's Guide to the Galaxy"? At one point you would ask the program to look into a room and it would respond that the room was empty. It took several repetitions before the program finally admitted there was something there. A simple piece of programming to emulate the "orneriness" of the sentient.

[logical muse]

I remember hearing about some Japanese robot-dog that when they asked it to shut off, it allegedly "didn't want to" and wouldn't shut off.

Does this mean it's sentient and is aware it's alive and doesn't want to be turned off?

Yes.

[INRM]

Now are you *sure* about that there logical muse? :-p

[PixyMisa]

Well, it's true, for certain definitions of "sentient", "aware", "alive" and "want".

[Mercutio]

My car doesn't want to start on cold mornings. My computer hates me. My refrigerator is toying with me.

Funny how, when we speak of sentience, it comes down to behavior.

[Hellbound]

My car doesn't want to start on cold mornings. My computer hates me. My refrigerator is toying with me.

Funny how, when we speak of sentience, it comes down to behavior.

I thought it was a matter of having both a subject and a verb?

Oh, sentience, nevermind...

[PixyMisa]

:)

It has to come down to behaviour, because behaviour is what we can observe.

But yes, your computer hates you.

[MWare]

Clearly, Johnny Five is alive. Steve Guttenberg's career, not so much.

[UserGoogol]

The way I see it, consciousness is the degree to which a system "understands" itself. Thus, a post-it note with "please don't throw me away" is slightly more conscious than a brick, (because a brick has absolutely no self-referential properties, whereas the post-it note has some) but still a huge distance from the rather sophisticated sort of self-sentience that a human being has. Although it would depend on the particulars of the case, I strongly suspect that robotic dogs as they currently exist are much closer to self-referential post-it notes than they are to people. Simply saying "Please don't turn me off" does not require a particularly elaborate concept of self.

[baron]

Just because it's in the shape of a dog doesn't make it more prone to sentient behaviour.

Even if increasing complexity does eventually result in consciousness (and there's no proof it would) there are millions of vastly more complex pieces of computer hardware out there that would make the switch first.

[UserGoogol]

The barrier to sentient computers (whether in robots or otherwise) is that all useful computers are still von Neumann machines or variations thereof. In other words, they are completely algorithmic. They will ultimately only react the way they are programmed to react (even if that algoritm includes some virtually unpredictable randomization routine).

Of course, one may philosophically claim that humans (which we define as sentient) are also ultimately algorithmic.

Hans

And even if humans aren't algorithmic, we're still "biomechanical," and that's not really that much better. Even if our behavior cannot be reduced to some sort of Turing-computable algorithm, we're still just the aggregate behavior of a bunch of complex (and not so complex) molecules behaving like complex molecules, which is really more plausible to have free will than a Turing machine.

Also, the "react the way they were programmed to react" argument bugs me for other reasons. It's fairly common that programmers can create a program that does not behave the way they intended it to, both in the obvious case of bugs but also the more interesting case of an algorithm having emergent but beneficial properties one might not expect. And if "how they were programmed to act" does not mean "how they were intended to act" then it simply means "Programs behave the way their nature makes them behave," which is kind of true of all things.

[logical muse]

Now are you *sure* about that there logical muse? :-p

I can't see how you could come to any other conclusion.

:p

[rockoon]

Its the path of least resistance, just like everything else that ever was, is, or will be.

[INRM]

Logical Muse,

How do I have this feeling that you're being somewhat sarcastic?

[crackers]

Logical Muse,

How do I have this feeling that you're being somewhat sarcastic?

You were programmed to have that feeling.

[wuschel]

I remember hearing about some Japanese robot-dog that when they asked it to shut off, it allegedly "didn't want to" and wouldn't shut off.

Does this mean it's sentient and is aware it's alive and doesn't want to be turned off?I remember hearing about some Japanese motor-vehicle that when they asked it to start up, it allegedly "didn't want to" and wouldn't start up.

Does this mean it's sentient and is aware it's dead and doesn't want to be reanimated?

[logical muse]

You were programmed to have that feeling.

:wackybiglaugh:

[logical muse]

Logical Muse,

How do I have this feeling that you're being somewhat sarcastic?

Busted...