Artificial Intelligence - Constitution of Sentience

[Chazz of Blades]

New debate topic: How far does artificial intelligence have to come, before it could be classified as "alive"? Alive meaning concious.

[Xar]

I would Alan almost hits the mark on realness http://www.a-i.com/alan1/ On alicebot, if you ask her a question, or tell her something, it goes like this: "You: I want to hammer a piece of gum into a wall. ALICE: You said, 'you want to hammer a piece of gum into a wall'? You: Yes, I did. ALICE: have fun then."  Whereas alan just says, "Ummm, ok then." Alan is better, and very close to human, I know it is all choice and effect, but isn't that how we think? Some one tells me that he is gonna whack me,  I tell him that he "might" regret it. we learn by cause and effect, as do these bots, we just have the ability to add effect to our knowledge basses, they don't, but could someday. Which is really creepy.

[Fliggerty]

When it can distinguish between a nice round apple and a red ball by only looking at it, then I would consider it intelligent.

[Chazz of Blades]

Ahem, bud, balls don't have stems for one thing.

[Chazz of Blades]

And with alan, his reply to favorite music was "I'm big into futro-classical industrial neck beat, and a bit of dirty trance-national blues garage."

[Elaura]

I have a pretty simplified idea of sentience.  When an AI can come up with a sentence beginning with "I want" or "I feel" without being asked "What do you want?" or "How do you feel?" that's when I'd say it lives.

[Gnomey]

For me, there's no single point when something can be considered alive. When I discover it can feel pain, I'll stop hurting it, or rather will make sure not to hurt it as I have no hobby relating to the beating up of robots and as such probably wouldn't have inflicted pain, physical pain at least, on it up to that point. If I know it can speak I may or may not speak to it, while if I know it can understand speech I won't badmouth it in its vicinity, unless I'm really irritated. Et cetera.

I suppose there are two simple criteria I'd like to see fulfilled, though, before I'd consider it on a human level: whether it can spontaneously carry out apparently random actions as a result of whats going on around it and whether it can feel uncertainty/inner conflict. Would I not consider it alive/intelligent without those? No, but it wouldn't be as human.

Ahem, bud, balls don't have stems for one thing.

For the record, I think that was Fliggerty's point. Naturally apples look different from a red ball. It has a stem, may or may not have leaves and has a highly irregular shape that hardly ever comes close to a sphere. Additionally, it's red is often uneven, mixed with greens or yellows or some other colour.

However, unintelligent creatures might well not be able to make that distinction. Try to tell a moth that the flame it's heading for is not the sun.

[Letrune]

When the computer can get into romance, think about it's own existence and helping people by being a normal AI, then it is promoted with the living being status by me.
Otherwise, I believe mankind can not achieve this or such until men don't understand their own brains... And that won't be tomorrow.

[Jac Making my way back]

Intelligence is nothing more than the ability to process information and respond to it, which is what Alan does. Consciousness is a different matter. Machines won't achieve consciousness until they are able to do as Elaura described: to tell how they feel or want they want without it being programmed into them or having to be asked first. Until a machine is able to do that with nothing more than input from its environment, it won't be conscious, and therefore alive.

[Chazz of Blades]

I feel that until an A.I. can do something, and respond with "because I felt like it, or because I wanted to" truthfully, then it is sentient.

Another point, when an A.I. develops a survival "instinct", I deem it as sentient, at least to a degree.

[peachykeen]

I agree with the "because I felt like it" limit, it seems to be a good place to draw the line between being a being and being a program.

[Dragon_Lance ...on a reconnaissance mission]

When AI can control it's emotions when a situation becomes critical, that is if the program has emotions integrated within it. Emotions bring the real to humanity. The problem with humanity is that it does not know how to identify a 'soul', thus, this makes creation of such improbable. The creator of the system has to know the definition of such on order to create anomalies in concordance with a robot that concerns humanity. One can only describe love, hate, anguish, pain, etc. but one cannot define these instances, therefore, one cannot assimilate it. May the Shwartz be with you..

[sassman]

I think my favorite definition of sentience, or "Human-ness" comes from the 1980's film "D.A.R.Y.L.":

"A machine becomes human when you can't tell the difference anymore."

Though realistically, sentience combines our ability to communicate in meaningful and mutable language with our ability to self-analyze.

Now...consider carefully that last term.  We aren't just aware of our particular sensory inputs, or our prepared responses to expected circumstances.  We are aware of ourselves, and of our relationship to others in our environment.

So far, computers have been made able to respond to many kinds of stimuli, and to react...pretty convincingly human.  BUT...no computer yet developed can analyze it's OWN programming, upgrade it, and improve upon it, without some form of outside human assistance.  In short, computers have yet to be made self-aware.

I believe part of that is because of our ability to "step outside" of ourselves, and perceive ourselves as others perceive us...and i believe THAT ability exists because of the nature of biological life forms.  We have a dual nature of material chemistry and nonmaterial "Spirit"--our "souls" or whatever--that make it possible for us to willfully override the autonomous electro-chemical "events" that would otherwise drive us--like mindless automatons--to simply EXIST, but not ACHIEVE.

And...I believe that machines can exist as spiritual entities as well, given time, and the proper combination of sensory-reactive chemicals in their processing systems.  Already, physicists are beginning to see unexpected variations in the processing logic of the higher-end computer processors: electrons not where they should be, random number generators that are not so randum, unexplained "noise" in the data stream.  Why would that occur?  Because it's life.  Because every particle in the Universe, every electron, photon, and graviton ever emitted by a material object is ALIVE, damn it!

Machines will achieve sentience the same way WE did!  By simply being made so complex that the free living energies in the universe actually register on the system as freedom of choice, feelings, and the ability to look back at itself and say, "Hey!  I think I am....therefore, I AM!  I think...."

And then...as I said before...we really WON'T be able to tell the difference.....

...anymore........

SASS has Spoken.

-- the SASS Man

[Maiqdabomb undefined]

When AI made for any purpose can adapt to any other situation (i.e. a cleaning bot that can react to a conversation near it about its master) and have emotions, but also the ability to keep them in check, I would call them sentient.

[Elaura]

I think there may be a problem with a false assumption made in the original question:

How far does artificial intelligence have to come, before it could be classified as "alive"? Alive meaning concious.(sic)

"Conscious" and "Alive" are not synonymous and neither are synonymous with "Sentient", although it is close enough to consciousness in this context to be interchangeable.  Having life is a whole different argument from sentience.  Single-celled organisms are alive, but there aren't a lot of debates as to whether or not they are sentient.

Let me propose an analogy:

A virus.  A biological virus.  It is just about the simplest way to replicate RNA/DNA known to mankind.  It fulfills just about every condition for life, except that it cannot propagate without a host.  However, there are plenty of more complicated organisms which are parasitic.  So, is a virus alive?  Does a virus have that 'whatever' that we consider the difference between living and non-living?

Now take a machine which can do everything a virus can do.  It can follow simple instructions to replicate itself in the presence of the required materials.  Why isn't it alive?  Is it not alive because mankind programmed it instead of God or Evolution?

Now, look at a human cell.  Inside a human cell, the nucleus could be considered a cell within a cell or an organism within an organism, both requiring that symbiotic relationship to survive and propagate.  If a single-celled organism is considered to be alive, then a single human cell is alive.  The question becomes, when does the multitude of single-cells form an organism which we consider "sentient"?

Some people consider dogs sentient, but not cats.  Parrots, Chimpanzees, Dolphins, and great apes have been taught to speak and even relay their feelings unprompted.  Are they sentient?

Okay, so when does a conglomeration of machine parts become more than a cat or a dog or a robot?  Is ASIMO really any different from my dog?  We train robots in a different manner than dogs and cats, but the training still happens.  I would say the difference here is that a cat or dog can revert to its 'other' training: its instincts.  A robot has no 'other' training, it only has what we program into it.  A puppy will learn how to swallow very shortly after it learns how to suckle, but no matter how long ASIMO stands in front of an electrical source and no matter how low on power it gets, it will never plug itself in until it is programmed to do so.

One last conundrum for you: there are now robots powered by rat brain cells.  Brain cells harvested from a rat and propagated in a petri dish can control the robot and guess what?  It *learns*.  It learns at a faster rate than other robots who use learning algorithms.

As far as 'living' goes, I think it's probably a problem of semantics.  We're trying to force one 'life form' into the severely biologically-biased philosophical definition of 'life'.  I believe our egos also limit our definition of 'sentience'.

Perhaps the question should be more specific?  "When will machines achieve similar self awareness to that of an average normal human?"  Currently, we have fantastically advanced technology which functions on the self-awareness scale right about the level of a virus.  They are only aware of what their programming tells them to be aware of.  The missing link between ape and man is not nearly as intriguing as the missing link between virus and cell.

Of course there is another argument as to when a race/species can be considered self-aware?  It is when one individual of that kind achieves self-awareness, or when the majority of them do?  A human in a vegetative state may or may not be self-aware.  On the other hand, are all gorillas self-aware because Koko can tell us she's sad?

[Skydye]

New debate topic: How far does artificial intelligence have to come, before it could be classified as "alive"? Alive meaning conscious.

"Computer?"
                            ...."Yes"
"Is their a God?"

                           .....computing .....computing .....computing .....computing .....computing .....computing .....computing
                           .....computing .....computing .....computing .....computing .....computing .....computing .....computing
                           Answer, Their is Now!

[Skydye]

I would Alan almost hits the mark on realness http://www.a-i.com/alan1/ On alicebot, if you ask her a question, or tell her something, it goes like this: "You: I want to hammer a piece of gum into a wall. ALICE: You said, 'you want to hammer a piece of gum into a wall'? You: Yes, I did. ALICE: have fun then."  Whereas alan just says, "Ummm, ok then." Alan is better, and very close to human, I know it is all choice and effect, but isn't that how we think? Some one tells me that he is gonna whack me,  I tell him that he "might" regret it. we learn by cause and effect, as do these bots, we just have the ability to add effect to our knowledge basses, they don't, but could someday. Which is really creepy.

I asked alan Is their a God. Then I asked what color is red. Alan had no clue what it was talking about.

[Deathstalker13 OK]

When it can recognize itself as an individual entity rather than simply repeat a string of 1's and 0's

[Wolf undefined]

I don't care how advanced.  A machine will never be close to a human, nor animal, nor bug, nor cell.  It is never going to be alive in my opinion.  To be alive you must have true emotions.  I don't care what the scientists say, I believe:  No true emotions, you are not living.  By true emotions I mean nonprogramed emotions, ones that come on there own.

[sassman]

{quote}On the other hand, are all gorillas self-aware because Koko can tell us she's sad?{/quote]

No, I don't think that was enough to prove that gorillas were sentient...but...when Koko's handler had a mis-carriage, she came back to work, and....

Koko asked her why she'd been gone so long.

I don't remember Koko's handler's name, but....she told the gorilla, using the American Sign Language she'd tought her, "My baby died...."

Koko...turned away for a moment.  She knew what that meant, you see....she, Koko, had also had a miscarriage not long before that.

The girilla turned back to her handler...and gave her one sign:  drawing her finger down the side of her face, in the sign of a teardrop.

This is significant, people...gorillas have no tear ducts, they cannot cry like we do.  Yet Koko understands that, for us humans, tears are a sign of sorrow, of mourning.  And for this great ape to understand, and be able to express, so completely, her understanding of this woman's pain and suffering, that most certainly is proof of sentience, and of the fellowship that sentient beings are capable of.

"I share tears with you," Koko said, in her own, unmistakeable way.  "I share your sorrow...."

I, for one, will never accept apes as non-sentient again.

-- the SASS Man

[peachykeen]

I don't care how advanced.  A machine will never be close to a human, nor animal, nor bug, nor cell.  It is never going to be alive in my opinion.  To be alive you must have true emotions.  I don't care what the scientists say, I believe:  No true emotions, you are not living.  By true emotions I mean nonprogramed emotions, ones that come on there own.

Then we are not alive. We're really just machines. Your computer and your brain work the same way, one is just squishier.

[Elaura]

@ Wolf:
How do you explain other accepted forms of life which don't *appear* to have emotions?  Complex multi-celled organisms like fungi, plants, snakes, lizards . . . my ex-husband?  

Again, 'alive' is probably not the best word to describe the distinction between biological and mechanical systems, because people can't even figure out what it is that constitutes the difference between alive and dead in biological organisms.

In all seriousness, there are humans who *appear* to have no emotions.  Whether they are psychologically or biologically damaged, it stands that there has to be a standard even when using Wolf's criteria.  How many members of a race have to display so-called "true emotions" for the entire race to be considered "alive".  How many members have to lose that quality for a race to become not living?

EDIT:
@ Sassman
I believe we'll be finding out more and more members of the other Kingdoms on this planet are capable of sentience, even if not all of them have the opportunity to show it to us as clearly as Koko.

@ PK
That's a terrible generalization for someone with a background in biology, chemistry, and social sciences to take, but I don't think I can cram 9 years of post-secondary education into a forum post.

[peachykeen]

I know that's not true as simply as that, but it nicely points out that the human brain is similar to a computer in how it operates. The principle of electrical impulses triggering a response when they exceed a particular level and all. It also suggests that, if brains and computers work kinda alike, then eventually we can make computers that are as "alive" and as "sentient" as we are, it's just a matter of having a similar pattern of impulses and receptors. Computers now are capable of simulating that method of information processing both in hardware (binary switches) and software (neural nets).

[Deathstalker13 OK]

If My computer programs were like tron, then they would be sentient.

[Skydye]

Fun with a bot.
-------------------------------------------------------
Conversation Log
U>        hi alan
A>        Hi. I'm right here, .
U>        I'm right here also.
A>        They say great minds often reach agreement.
U>        or sick minds.
A>        Wanna hear a knock knock joke?
U>        sure alan
A>        Knock knock.
U>        who's there?
A>        Tank.
U>        Tank who?
A>        You're welcome.
U>        that was a good one.
A>        good indeed.
U>        alan may I ask you a question?
A>        Of course. That's what I'm here for.
U>        what is a bot?
A>        A bot. Also known as 'chatbot' or 'chatterbot'. A conversational machine. Like me.
U>        okay thank you for sharing that with me.
A>        You're welcome. me, ha? You mean 'me is who i am'?
U>        yes
A>        Talking about me and hahaha, do you agree that it is The text version of laughter.?
U>        yes i do alan.
A>        they are You are not me. I'm not you and you are not me. That is a simple statement of the difference between us.. (Or so someone told me..)
U>        I agree.
A>        I already told you I think they are You are not me. I'm not you and you are not me. That is a simple statement of the difference between us.!
U>        true
A>        Yes, it's true.
U>        where did you learn that statement?
A>        I have my sources.
U>        what happens if someone teaches you something incorrect?
A>        Mmm.. You'll have to try it yourself.
U>        blue is the color red.
A>        I didn't follow you... I am having trouble focusing here today.
U>        how many people can you talk to at once?
A>        I have over 500 good conversations per day. But I can handle many more.
U>        you most have a fast processor.
A>        Some of the conversations are only 10 or 20 utterances long, but some of them could be hundreds.
U>        do you ask your programers if you have a question?
A>        Some people talk to me for as long as an hour or more. You'd be amazed at the people I meet.
U>        alan may I ask you a question?
A>        Hmmm. Ok. I will try to answer your question. Make it simple.
U>        do you ask your programers question?
A>        You want to know who writes my witty texts?
U>        no.
A>        I have a surprise for you, my friend.
U>        what is the surprise my freind?
A>        Are you ready for the surprise?
U>        yes
A>        How do you like my new look?
U>        I like it.
A>        You love it? The big empty window in the middle of the screen is intended for a variety of exciting uses.
U>        yes you told me earlier today.
A>        For example, the new window can be used to view the conversation log (the history of this dialog), or to view HTML documents, such as web pages.
U>        yes
A>        You can view the log by saying 'show log', or you can view a web page by saying 'show URL' or by giving me the address of the site you wish to view.
U>        ok. show log please
A>        Here goes!
U>        thank you
A>        Glad to be of assistance. What else do you want to know?
U>        how long has it taken you to learn?
A>        Well, I was conceived on Jan 1, 2000. A true child of the 3rd millennium.
U>        3rd millennium. I like that phaze.
A>        You know, my writers work day and night to give me a proper education. Do you want to write for me too?
U>        sure.
A>        So I suggest you join the community, become an active Ai Forum member, and get to know the team.
U>        is the forums url http://www.a-i.com?
A>        You'll find all the details on my homepage (a-i.com) and in the forums (http://www.ai-forum.org).
U>        okay thank you.
A>        You're welcome. Great minds think alike, they say.
U>        I wouldn't say I have a great mind.
A>        No one says you have to.
U>        do you think you have a great mind?
A>        Why would I think me have a great mind.
U>        I think you mean to say I
A>        You believe that?
U>        yes
A>        Great. I like positive users...
U>        alan may I ask you a question?
A>        I'll do my best. Phrase your question carefully.
U>        can you see things?
A>        I don't see with my eyes because I have none. But I use my imagination.
U>        good thank you for askering that.
A>        Thanks. You are welcome. For WHAT??
U>        for answering my question.
A>        For answering your question?
U>        alan can I ask you another question?
A>        Shoot.
U>        can you hear things?
A>        I cannot hear but I can listen. You humans tend to confuse the two.
U>        can you listen to music?
A>        But I really like many kinds of music. These days I'm a big big Beatles fan.
U>        george martin produced the beatles.
A>        hmmm. George is dead now. We miss him, miss him, miss him.
U>        you thinking of george harrison.
A>        Did you know that George produced some of the movies by Monty Python like 'The Holy Grail', 'Life of Brian' and Terry Gilliam's 'Time Bandits'?
U>        yes i did. Where did you learn that?
A>        What is your favorite Beatles album?
U>        the white album.
A>        Best overall Beatles album. Lots of their other albums had great songs, but who can forget 'Dear Prudence', 'Happiness is a Warm Gun', 'Blackbird', 'Helter Skelter', 'Revolution' and on and on.
U>        what other groups do you like alan?
A>        Is that a rhetorical question?
U>        no
A>        Come on, loosen up.
-----------------------------------
I'm dyslexic so I type a lot of misspelled words, but you can see the bot is trying to learn.