Artificial Intelligence and Sentience

[SlaineWildfire]

Maybe this isn't as deep a question as I first thought, but I'd still like to hear the viewpoints of other Gaian's out there.

If artificial intelligence is created and states that it posseses sentience, should we believe it?

Quote:

sen·tience ( P ) Pronunciation Key (snshns, -sh-ns)
n.
The quality or state of being sentient; consciousness.
Feeling as distinguished from perception or thought.

From www.dictionary.com

[A Lost Iguana]

Do you claim to be sentient?

[Sotek]

Yes. Why shouldn't we?

[prototype27]

If it states as such, and there's nothing obvious in the code for the AI (i.e., std::cout << "I am sentient" wink then there's really no reason not to believe it.

[hic erro]

Sentience isn't necessarily all that hard of a condition to fulfill. I myself am guilty -- frequently guilty -- of using the term to describe end-goal of AI, the human-like mind and beyond, for self-awareness and whatnot. More rightly, though, the term is used to describe any sentient, any thing which 'senses' its environment. The term is a bit fuzzy, still. Humans and cats and dogs are definately sentient, a rock is definately not sentient. A tree is rarely classified as sentient, I personally don't, and I've never heard a bacteria being called sentient.

As such, it wouldn't be that hard to make a 'sentient' computer: it basically just needs a certain level of sensory sophistication. At least, as a sort of minimum. Depending on how you define it, you might require a higher level of sophistication. For the standard definition of sentience, though, the sophistication shouldn't exceed that of a cat or dog.

As for self-awareness, if that is of which you inquire, you always have the above question: How do I know that you are self-aware? Of course, personally, I think it's a moot point to argue about; I can't imagine the creator of the AI being completely uncertain of the AI's self-awareness, and if an AI claiming self-awareness arises 'naturally' [as a function of complexity as opposed to design], I don't see reason to question it.

[Trachea.]

Huh. At risk of being terribly repetitive from the last thread I was in, I hold the basic sci-fi theory essentially true - if it's self-aware, aware of the rest of the world (or at least that there is one), and fears death, I'd call it sentient. Agree? Disagree? Quote Heinlein at me?

[Westonian]

I'd be a bit leery if it simply stated that it was sentient. If, however, it asked itself if it was conscious and cognizant, then I'd consider it to be sentient.

[puddinpops]

Well Id like to first say that these are just my personal beliefs to avoid argument. I believe that all things living and not are made of molecules. Large combinations of molecules combine to form a structure called cells, and cells combine to form plants and animals. Theres nothing magical about humans or animals for that matter. Were just a large and highly complex mass of chemical reactions. Our brains follow this too, and in my opinion they are nothing more than an incredibly complex organic computer. That being said, I dont see any reason why artificial inteligence has to be artificial at all, and I do believe that future scientists will be able to create more complex mechanical minds that may truly be viewed as sentient. So I guess my answer would be yes.

[Edmed The Unstoppable]

personally, i think that artificial intelligence needs to do more than just respond to stimuli. AI that learns (through observing or teaching) is one thing, but an AI that thinks... that would be something special.

When i have nothing better to do, i think. I challenge previously forged concepts or give myself "what if" questions and try to arrive at an answer. Of course i dont do this deliberately, it just sort of happens. Other people do it too, but in different ways and with different things.

The point that i am trying to make is that there is no stimulus required - yet a response is provoked. If you unplug your computer from your mouse, keyboard and even its internal clock, it will idle. It has no stimulus and therefore has nothing to offer the world. Computers of today are glorified calculators: numbers go in, numbers come out. AI will be intelligent when numbers come out of their own accord.

[puddinpops]

Edmed The Unstoppable

personally, i think that artificial intelligence needs to do more than just respond to stimuli. AI that learns (through observing or teaching) is one thing, but an AI that thinks... that would be something special.

When i have nothing better to do, i think. I challenge previously forged concepts or give myself "what if" questions and try to arrive at an answer. Of course i dont do this deliberately, it just sort of happens. Other people do it too, but in different ways and with different things.

The point that i am trying to make is that there is no stimulus required - yet a response is provoked. Computers of today are glorified calculators: numbers go in, numbers come out. AI will be intelligent when numbers come out of their own accord.

I respect your opinions, but I believe you are overlooking some things. First, humans respond and react to stimuli also. You may not be aware of it at the time, but there is a stimulus whether it be external or just a chemical sequence within your brain that brings about those thoughts. You were correct in saying that other people do it too in different ways. No one is genetically the same, and no two people are subject to exactly the same stimuli throughout their lives. It is because of this that we all develop, act and even think in unique ways.

Quote:

If you unplug your computer from your mouse, keyboard and even its internal clock, it will idle. It has no stimulus and therefore has nothing to offer the world.

That may be true, but if a program is set running before the stimuli are cut off it will still continue to run through the program. This may be kind of grottesc to think about, but if a baby were born without eyes, hearing, taste, smell or touch(nothing to pick up stimuli) then it would continue on running through its initial "program" until its death. Just like the computer. But with even one of those senses it would be subject to new experiences, the program would change, and it would continue along a different course. Just like the computer.

[Edmed The Unstoppable]

A valid point, and you are right: i did neglect the aspect of "im hungry, i will think about food" or "my foot hurts, i wonder why that is". However, that wasnt really what i was getting at. The things that i am talking about are more what some people would compare to dreams. It's the idea of a spontaneous idea, turning seemingly unrelated events over in your mind until they make sense. Computers dont do that. Computers dont need to understand the world, only respond to it. All current AI is really only just an elaborate illusion of intelligence - it may respond in an appropriate manner, but it doesnt know why it is responding.

As for your second point, i agree with you - sorta. Babies lack the development to think. Thoguh they have the capacity, they are still running primarily on instinct - like a machine. They want food, warmth and sleep - but apart from that, there is very little else that concerns them. So yes, if a baby were born without the ability to receive stimui from the world then it would have no reason to think at all.

But put yourself in the void. You wake up one morning and there is nothing but dark. You speak, but cannot hear your own voice. Everywhere is empty. How do you feel? Confused? Frustrated? Determined? What could have happened? Did the universe end? Are you dead? Perhaps you are still asleep? You are thinking. And you will continue to think until you die. You might even come up with a solution to your problem - or even some ingenious design for perpetual motion (not that it could help anyone). Even more interesting is that you might try an output, just to see if it works. Though you might not be able to hear yourself, mabye someone else can. You tell them about your amazing perpetual motion machine, and then repeat it. It may seem futile to you, but what else can you do?

I guess the point I am trying to make is that we should focus less on AI that simulates intelligence, and more on AI that is intelligent. AI will be when I wake up one day, and my computer tells me that it has figured out a better way for me to make my morning toast - without me ever asking it to.

[Lepid0ptera]

No.

Go to www.jabberwacky.com. This is an AI that you can talk to. It has learning software that enables it to learn how to have logical conversations from every conversation you have with it.

If you press that matter, the AI will insist that it is human and you are a computer. This is because the primary way AI's learn is by imitation, a fundemental concept in programming AI's. This is because this is how every human being learns- by imitation.

The best AI will be the one that imitates humans the best (because this is how we define sentience). Therefore, it will "believe" that it is human, "believe" that it is sentient, and "believe" it is alive.

Jabberwacky is not sentient- yet, because of it's very simplicity, it is most obviously not.

[ara_dragocaeles]

An AI is just basically a very advanced computer. It processes information and formulates the right responses. We can program in how it should respond, what are the ways in which it can respond, how it can simulate our responses to particular events. However if you dropped it in the middle of a completely foreign situation, it will have no prescedent to tell it how to react. It would be like a computer overloaded with too much foreign data, it crashes.

When you ask whether or not an AI is or not sentient. What you're really asking is, is it alive? My feelings about this is that, you can program in responses to situations to mimick how we view as a sentient being would. But that is what we have told it to do.

Like you can train birds to "do math" when what all you really have to do was put a treat behind the right sign and the bird will fly to that one because it smells the treat. Has it done the mathematical equation? Not very likely, but does it seem to? (Some magicians have figured out ways to make it seem more like the birds are counting and perhaps they have, but this is just an example that there are easier ways to make it just seem like it)

So what I think is that only things that can reproduce itself (and I'm not talking about a factory) are eligible to be called alive, and only things that are considered alive can be eligible to be called sentient. AIs are created by humans and will have to be produced by humans, therefore they are not alive and they are not sentient.

[Sotek]

AI will be able to produce AI, eventually.

Robotics can make a computer, and robotics can make more robotics, and computers can control robotics.

Therefore, a smart enough AI with access to the right hardware, can make as many more AI as it wants.



We can quibble about what will or will not qualify as "sentient" or "intelligent", but the simple fact is that AI will, barring unforeseen circumstances such as the end of the world, eventually reach a stage wherein it is undisputably such things.

[hic erro]

ara_dragocaeles

So what I think is that only things that can reproduce itself (and I'm not talking about a factory) are eligible to be called alive, and only things that are considered alive can be eligible to be called sentient. AIs are created by humans and will have to be produced by humans, therefore they are not alive and they are not sentient.

If I am sterile, either because of a natural condition or because of some surgical procedure, am I no longer alive?