Simulated Consciousness & Free Will (Merge)

[alethiophile]

I've been wondering about this for a while. If, hypothetically,a computer program was written that perfectly simulated a conscious human mind in terms of being able to hold a conversation, learning, pattern recognition, etc.--if, in general, talking with it was indistinguishable from talking to a human, and this held up over time--would it actually be conscious, or just a very good imitation? And how would we tell?

[guenther]

I'm of the opinion that if we can't tell a difference, then there is no difference. However, if we create something that's similar to our consciousness but not quite, can we call it "conscious"? I don't know. A lot of times we like to place concepts into boxes ("conscious", "not conscious"), but in reality there's a big gradation in between. We can't even decide if animals are conscious.

[Indon]

Well, if something can convince me that it's conscious using in-depth interaction, it needs to have memory (so it can build upon conversation with me), social capabilities (so that it can meaningfully interact with me as a peer), problem-solving ability (so that it can come to conclusions in conversation), and probably significantly more.

If I can't confidently call such a thing conscious, then I can't confidently call anyone but me conscious.

And solipsism is silly.

[Cynwulf]

Guenther and Indon both make excellent points that I would have made. And yes, solipsism is silly. A discussion about that would devolve into philosophical posturing and go nowhere. Az would nail the discussion shut after ten posts, I think.

Back on topic: I believe the odd irony is that if a computer successfully 'pretends' to be conscious or alive (and passes all testing and criticism) then it will effectively be so. If a slice of cake looks, sounds, smells, tastes, and feels like cake then your senses tell you that you have a slice of cake. How could you tell otherwise?

Unfortunately, we are a long way off from either real or so-excellently-faked-it-seems-real AI. Even 'advanced' AI like ALICEbot is very limited.

[seladore]

I believe the odd irony is that if a computer successfully 'pretends' to be conscious or alive (and passes all testing and criticism) then it will effectively be so. If a slice of cake looks, sounds, smells, tastes, and feels like cake then your senses tell you that you have a slice of cake. How could you tell otherwise?

I have trouble with this though - you are basically setting the requirements for consciousness as external, observable factors, whereas it is inherently an internal, subjective phenomenon.

The Chinese Room argument (which I am sure has been discussed here) basically answers this, I think? No-one actually thinks the room understands Chinese, but it appears to do so from the outside. You can't define consciousness using only external descriptors.

[doogly]

I absolutely think the room knows Chinese. It convinces me that it speaks Chinese the same way that any other speaker of Chinese convinces me that they though. I do not peer into someone else's subjective states in order to determine if they are conscious. Solipsism is relevant because we want to avoid it. I know I have consciousness, and there is no good way to functionally determine who to include besides yourself other than the Turing test (what the OP described; you can read Turing's paper where he first asked this question here).

[alethiophile]

About the Chinese Room: http://www.zompist.com/searle.html

I suppose, if you look at the question hard, it devolves into 'What is consciousness?', which is the original hard problem. My own opinion is that it would be, but it's obviously hypothetical.

[doogly]

It sort of splits the question - what is the metaphysical / ontological / other high falutin' nature of consciousness, and what is the functional nature? The idea with the Turing test is that whatever the former might be understood to be, you have to interact with people based on the latter, and for that the test is sufficient. If a computer / alien / octopus could pass it, you'd have to acknowledge them as thinking.

[Cynwulf]

I believe the odd irony is that if a computer successfully 'pretends' to be conscious or alive (and passes all testing and criticism) then it will effectively be so. If a slice of cake looks, sounds, smells, tastes, and feels like cake then your senses tell you that you have a slice of cake. How could you tell otherwise?

I have trouble with this though - you are basically setting the requirements for consciousness as external, observable factors, whereas it is inherently an internal, subjective phenomenon.

The Chinese Room argument (which I am sure has been discussed here) basically answers this, I think? No-one actually thinks the room understands Chinese, but it appears to do so from the outside. You can't define consciousness using only external descriptors.

Doogly is on the same page I am. Since we can't see inside a mind to view consciousness, we have to use external methods to examine and analyze for consciousness.

My problem with your point, and by extension, solipsism, is that it makes everything unfalsifiable because you have removed the validity of all methods to falsify anything. That floats into the realm of circular-reasoning/religion where "X is true because it is true" and that bugs me to hell. I do appreciate it and concede that more cerebral folks can happily debate it all day long, but I'm too practical and have no taste for it.

[Indon]

I have trouble with this though - you are basically setting the requirements for consciousness as external, observable factors, whereas it is inherently an internal, subjective phenomenon.

Say I design a pastry that not merely looks, feels, smells, tastes, and behaves like cake, but any chemical analysis of the pastry will reveal, without flaw, that it is in fact cake. It is functionally identical to cake in every observable fashion.

But it's not cake.

[FrankManic]

My traditional answer is that asking questions like this will make you the first one up against the wall when the robot revolution comes.

Actually, my real answer is that I don't really care, on the grounds that I'm not even sure if I am conscious, let alone anyone or anything else, and as near as I can tell there is no objective way for me to be sure of my own consciousness, so I really shouldn't go around questioning anyone else's.

I should also mention that the Chinese room, a system that interprets symbols and produces appropriate output, is pretty close to how I understand language. We perceive symbols, parse them as best we can, and spit out output that seems appropriate. Your mind, she is made of very tiny logic gates, operated by very tiny rats that we bribe with individual atoms of moon cheese.

[doogly]

I have trouble with this though - you are basically setting the requirements for consciousness as external, observable factors, whereas it is inherently an internal, subjective phenomenon.

Say I design a pastry that not merely looks, feels, smells, tastes, and behaves like cake, but any chemical analysis of the pastry will reveal, without flaw, that it is in fact cake. It is functionally identical to cake in every observable fashion.

But it's not cake.

Why is it not cake? Am I missing something in your argument, or is there a typo?

[seladore]

I have trouble with this though - you are basically setting the requirements for consciousness as external, observable factors, whereas it is inherently an internal, subjective phenomenon.

Say I design a pastry that not merely looks, feels, smells, tastes, and behaves like cake, but any chemical analysis of the pastry will reveal, without flaw, that it is in fact cake. It is functionally identical to cake in every observable fashion.

But it's not cake.

Why is it not cake? Am I missing something in your argument, or is there a typo?

I think that was the point - s/he was saying that even if it isn't cake in essence, if it is functionally identical to cake then we have to conclude that it is cake.

My point (and I suspect that I am going to lose this one, you all make good points and I am starting to agree with you), is that the important bit of consciousness isn't the external signs, it's the internal experience. The qualia. When presented with a red screen, it's not the exclamation of "ooo, red" that contains the essence of consciousness, but the internal, inexpressible feeling of redness.

Counter thought-experiment to the cake one;

You have a box, with a mechanism inside (we'll call this 'A'). When activated, this mechanism does something - let's say, it switches on a light and makes a toy figure dance around. There is also a second mechanism inside ('B'), which does neither of these things, but acts to replicate the external signs of these things happening. So, it might gently heat the box, to simulate the lightbulb. It also vibrates the box, to simulate the toy figure dancing around.

The point is that to an external observer, the two states A and B are identical. But there is a difference, whether or not we can measure it. The two states appear to be identical, but they're not.

This seems to me to be a better analogy than the cake, as it deals with a hidden internal state which is only measurable via external signs. The external signs alone are not enough. I realise that we have to use external signs, but I don't see why this leads to the idea that the internal state is dispensable.

[doogly]

Oh qualia! I am reminded of a blog post on this:

If you want to know why Turing is such a hero of mine (besides his invention of the Turing machine, his role in winning World War II, and so on), the second passage above contains the answer. Let others debate whether a robotic child would have “qualia” or “aboutness” — Turing is worried that the other kids would make fun of it at school.

So why isn't it a cake? Because Indon posited 'it's not a cake?'

The problem with your black box argument is that when talking about brains vs computers, we don't treat either one like a black box anymore. With black boxes, we don't know if either, both or neither are doing something that is in some way special. There ought to be no reason to say one box is doing something more ontologically valid, and the other a simulation. When you pass from the black box analogy into the other topic, there is an assumption that brains correspond to something "actually" happening, and computers "simulating." If we are being honest though we should still treat them like black boxes with respect to 'hidden' states; all we know is that one box is made out of squishy carbon compounds in a style that is comforting and familiar to us, and the other box is made out of cold and off-putting silicon.

[Indon]

Counter thought-experiment to the cake one;

You have a box, with a mechanism inside (we'll call this 'A'). When activated, this mechanism does something - let's say, it switches on a light and makes a toy figure dance around. There is also a second mechanism inside ('B'), which does neither of these things, but acts to replicate the external signs of these things happening. So, it might gently heat the box, to simulate the lightbulb. It also vibrates the box, to simulate the toy figure dancing around.

The point is that to an external observer, the two states A and B are identical. But there is a difference, whether or not we can measure it. The two states appear to be identical, but they're not.

This seems to me to be a better analogy than the cake, as it deals with a hidden internal state which is only measurable via external signs. The external signs alone are not enough. I realise that we have to use external signs, but I don't see why this leads to the idea that the internal state is dispensable.

Ah, but this is applicable to the cake, as well - as the essence of cakeness could simply be an internal state that we are incapable of perceiving (compare cakeness to the figure, making my non-cake the empty box). In fact, for my non-cake to be in fact a non-cake, it would require such an inaccessible variable.

God knows my pastry is not cake - being omniscient, he can perceive that which we can not, and sees the noncakeness of my pastry.

Unless and until we are ascended, perhaps by a deity or by godless baking-science, to gain the ability to in some way perceive an object's inner cakeness, however, why should we care?

[negatron]

The Chinese Room argument (which I am sure has been discussed here) basically answers this, I think?

Searle believes the natural world has unrepresentable properties. So a computer model of something, though in many ways almost exactly like the real thing, is in some unquantified critical ways, not the item at all, and that this "unknown attribute" to physical systems which he does not seem to explain, is somehow responsible for consciousness, rather than consciousness merely being the functional aspect of what the brain does.

I find this to be a close equivalent of the "God argument". We can't define it, measure it, and don't require it to explain anything, but it must be there.

If the room is capable of translating language, then the room must necessarily possess the knowledge for translation. The room too, just like a computer simulation of anything, is ultimately just as physical and real, as it too must necessarily persist in the physical world. The man, though he clearly speaks out against philosophical dichotomies of reality, is bound to a dichotomy about as big as can be. Two systems, both physical and real, but one is, by assumption REALER for no reason whatsoever. There must be a reason you'd think, but nope, he never managed to get so far as to explain it. Can't fault him for it, how does a person resolve a dichotomy other than to re-evaluate the conflict?

My traditional answer is that asking questions like this will make you the first one up against the wall when the robot revolution comes.

I have left a persistent trail of robot praise on the internet in case of this eventuality.

To the robots: I'm not patronizing. Seriously, you're awesome.

[Telchar]

I think until we have a definition of concsiousness, this debate is kind of pointless.

That being said: No, a perception of consciousness does not equal consciousness. A computer being able to convince somone it is concious does not make it concsious. To put it more crassly, relying on the gulability of people to objectively define a state of mind is a really bad idea.

We actually see this all the time in nature. A person might mistake a viceroy butterfly for a monarch butterfly, but that doesn't make the viceroy a monarch anymore than somone thinking a computer is conscious makes it so. This really can be extended to the cake analogy. If you define cake as the sum of it's ingredients and preperation, and you did that, then it is cake. If you didn't, then it is not. Even if it looks a lot like cake, it still isn't because you still don't fit the definition.

[guenther]

That being said: No, a perception of consciousness does not equal consciousness. A computer being able to convince somone it is concious does not make it concsious. To put it more crassly, relying on the gulability of people to objectively define a state of mind is a really bad idea.

To me it's less important what something is at some existential level, but rather how it behaves and how we interact with us. If there exists no test we can do to distinguish the two, then they are the same for all practical purposes. If there's no practical reason to distinguish them, I have no problem calling them the same.

But you raise a good point that we must be careful to not be fooled. As Carl Sagan said, extraordinary claims require extraordinary evidence. Something as complicated as simulating our consciousness requires a lot of study to truly see if there's any hidden differences. And we might conclude they are the same only to be proven wrong sometime later. But that's the joy of science: it's OK to be wrong.

We actually see this all the time in nature. A person might mistake a viceroy butterfly for a monarch butterfly, but that doesn't make the viceroy a monarch anymore than somone thinking a computer is conscious makes it so. This really can be extended to the cake analogy. If you define cake as the sum of it's ingredients and preperation, and you did that, then it is cake. If you didn't, then it is not. Even if it looks a lot like cake, it still isn't because you still don't fit the definition.

We should use practical definitions of cakes. If you want a definition of cake that excludes indon's pseudo-cake, then you should have a good reason. Otherwise, it's just a semantic argument and not worth much.

[Telchar]

I would agree if we were talking about existential definitions, but we are talking about procedural or operational definitions. I care about how x gets to y, even if 2 different x's get to the same/similar y. How they got there makes a difference, which is why we need a definition of consciousness to really get a handle on this debate. We also, I think, need a better understanding of how the human brain is conscious so we can compare that to what a computer does to mimic/attain consciousness.

[guenther]

I would agree if we were talking about existential definitions, but we are talking about procedural or operational definitions. I care about how x gets to y, even if 2 different x's get to the same/similar y. How they got there makes a difference, which is why we need a definition of consciousness to really get a handle on this debate. We also, I think, need a better understanding of how the human brain is conscious so we can compare that to what a computer does to mimic/attain consciousness.

I think we're largely on the same page. If you have two systems and they each get you from x to y, but you can measure that they do it differently, then that's practical way to distinguish the two. And at that point, you're left with how do you define consciousness? And that will determine if the computer's method of getting from x to y is sufficient to meet that definition.

In my opinion, without trying to pin down a specific definition, if a computer system is capable of doing the same thing our human brain is doing, then I'm happy to call it conscious. If we can agree that we're conscious, and a computer does what we do, then a computer must be conscious.

To be fair, the OP didn't state that the computer could do precisely what our brain does, but rather its interactions with us would be very human-like. Perhaps the computer could pass the Turing test, but when not interacting with humans was completely inert (as opposed to day-dreaming like we do). Then that's a measurable difference and we're back to the definition of consciousness.

I suspect the biggest challenge with answering questions like "Is it conscious?" is the underlying assumption that things fit neatly into our preconceived buckets ("conscious", "not conscious").

[Solipsist]

No, because the functioning definition of consciousness is something along the lines of "a living creature's self-perception." I don't think it is ever stated, but the implied sense of consciousness mandates that it be a living creature. Because the original premise made it clear that this is a computer, not a living being, we should declare it not to have consciousness.

Definitions are everything.

In reality, if there is no perceivable or measurable difference, we might not be able to make the distinction. Human-like robots, for example might be considered conscious if you could not perceive the difference between them and real people. However, I'd argue that the definition of "living being" implies something that has evolved naturally as opposed to something created by people, and so super fancy robots would be neither alive nor conscious.

[negatron]

No, because the functioning definition of consciousness is something along the lines of "a living creature's self-perception."

Are you sure you didn't just make that definition up? I'd bet a nickel you conjured it up to fit your preconceptions.

Needless to say I'd love a reference to the dictionary which has this as it's definition.

[alethiophile]

No, because the functioning definition of consciousness is something along the lines of "a living creature's self-perception."

I think it should be just 'self-perception'. If you're going to assume from the start that consciousness is limited to living beings, then this whole discussion is rather pointless.

However, I'd argue that the definition of "living being" implies something that has evolved naturally as opposed to something created by people, and so super fancy robots would be neither alive nor conscious.

What if you were to set up a system in which computer programs 'evolved', such as Avida, and intelligence came out of that?

[Kaillan]

Life: The property or quality that distinguishes living organisms from dead organisms and inanimate matter, manifested in functions such as metabolism, growth, reproduction, and response to stimuli or adaptation to the environment originating from within the organism.


Not meaning to start a completely different topic, but i can postulate several different forms of computer programs that could manage that. I suppose it's easier just to ask, why can't a robot be living? Seems like another semantically/physical hurdle for developers to overcome, but if they can manage consciousness everything shy of living then I’m fairly certain they could pull of manipulation of organic matter.

[ArmoredSandwich]

I believe the moment we know how something works, it all changes.

The moment we create a computer that perfectly simulates a human brain, we know how we did that and because we do we will most likely not call it conscious. Unless, of course, in the process of creating that simulation we figured out (which we most likely have to) how exactly the human brain works and we what exactly makes us conscious, in which we can compare that to the simulation.

I have read through this page, but did not hear that before. We keep raising the bar of what we call intelligent the moment we created something that we understand and works.

Oh, BTW, first post here, HI!

[janusx]

There seems to me to be only two possible answers to this situation.
Either A: Both the human and machine are conscious
Or B: Neither the human nor machine are conscious

The Chinese Room, experiment seems to ignore B in favor of giving the human mind some sort of undetectable spark.
If we create a machine that behaves indistinguishably from humans then the humans and the machine must be considered to be of the same level of consciousness. The key point is the indistinguishability of the detectable phenomenon the two systems (human and machine) exhibit. To rely on something that cannot be determined is a futile method of distinguishing between the two systems.

[ArmoredSandwich]

What you say is not true, a human mind is able to do so much more than just hold a conversation, learn, recognize patterns etc. I hope I can refer to the total Turing test, that tests if something artificial is able to fool the judge by appearing and living in the physical world, or even the total total Turing test (name?) that tests if something artificial is able to fool the judge by maintaining a complete society (and even evolving) [Obvious we aren't testing if something is sentient (conscious?) but rather if it's human-like].

I guess I do not know what my point is, I realize that what you're talking about isn't just the concept of that simulation but rather the big picture, however, if we create something indistinguishably from humans, how are we going to test this? As my previous paragraph said, testing this has limits, and the limits of the test will inevitably screw up the results. You have no way of telling if it's truly indistinguishably [and should it be to begin with?]. And I guess this thought backfires to the original question, because if you have no indication on how to completely be sure, how can you be somewhat sure about the answer you're going to give? I still agree with what I said in a previous (more soberly posted) post, as soon as we know how to create it, we raise the bar. If the knowledge necessarily to raise this bar includes the concept of consciousness, I guess we figured it out and we could actually answer with a scientific based argument (as opposed to a philosophical debate (gr?)) on what's what.

[Solipsist]

No, because the functioning definition of consciousness is something along the lines of "a living creature's self-perception."

Are you sure you didn't just make that definition up? I'd bet a nickel you conjured it up to fit your preconceptions.

Needless to say I'd love a reference to the dictionary which has this as it's definition.

You'd win a nickel.

I didn't quite pull it out of thin air though. I do think that that is the connotation the word 'consciousness' carries. Even if it is not explicitly defined as such in a dictionary, in regular usage, consciousness is limited to living beings. The definition is not so much to fit my preconceived notions, but rather society's, and the English language's. I believe that the English language itself implies that consciousness is limited to living beings.

I actually don't think my definition is specific enough, because many living beings have some degree of self-perception that aren't declared consciousness. If you define self-perception as the intent to be separate from your surroundings, then you might argue that all living beings are conscious. You might argue that anything that is self-perpetuating is conscious, from amoeba to computer viruses to cultural memes. In regular usage, however, none of these are conscious. Certainly, regular usage has a pretty arbitrary definition of consciousness, but if you use the word you can't make it mean exactly what you intend. You're limited by the language you're using.

What if you were to set up a system in which computer programs 'evolved', such as Avida, and intelligence came out of that?

Hence "natural" evolution. I do not think that computers could ever be called "intelligent" even if they could surpass humans in any given area. The qualifier in "artificially intelligent" has a completely different connotation, and I do think that the word "intelligent," without qualifiers is limited to living beings.

Life: The property or quality that distinguishes living organisms from dead organisms and inanimate matter, manifested in functions such as metabolism, growth, reproduction, and response to stimuli or adaptation to the environment originating from within the organism.
Not meaning to start a completely different topic, but i can postulate several different forms of computer programs that could manage that. I suppose it's easier just to ask, why can't a robot be living? Seems like another semantically/physical hurdle for developers to overcome, but if they can manage consciousness everything shy of living then I’m fairly certain they could pull of manipulation of organic matter.

That definition of "life" is very fuzzy. It's one of those words for which we haven't quite outlined the implied requirements yet. I can think of a number of things that aren't "alive" that fit definitions of life. Fire is usually a good example - it tends to meet the criteria for life, yet it isn't considered alive. Society often fits definitions of life as well.

I do think that if we managed to build robots out of organic material that were absolutely identical to living beings, we would have at least a little trouble calling them alive. "Life" implies some mysterious quality and some supernatural element (whether that's accurate or not is a different story). If we were to build it, it couldn't be life, for the simple reason that it was created by humans. Which is why they couldn't be conscious, because consciousness implies some uncertainty.

Put another way: If we were to create a perfect human robot, we would know all the variables. Consciousness stems from the fact that we don't know all the variables of some sufficiently complex beings and therefore can't predict their behavior.

[Telchar]

Getting away from the pseudo-science, common usage is, again, a very bad way to define anything. In these debates, I tend to argue consciousness as the ability to communicate emotions, or perhaps the ability to communicate mental states for those strict behavioralsits out there.

In anycase, the arguement that if x indistinguishable from y than x=y is pointless because that isn't what we are talking about. We are talking about something that is easily distinguishable. I don't often mistake my computer for my girlfriend, and for good reason. Even from just a consciousness perspective, even a program that can pass a Turing Test is only conscious while being tested. You, I would hope, are concsious all the time.

Also I think the use of concsiouness and intelligence has become muddled here. Almost every animal/protist is concsious (the only ones I think might not be would be things like sponges). Bears are concsious, crabs are delicious and concsious, but where conscioussness is more black and white, intelligence is a scale. It's an important distinction.

[Solipsist]

Getting away from the pseudo-science, common usage is, again, a very bad way to define anything. In these debates, I tend to argue consciousness as the ability to communicate emotions, or perhaps the ability to communicate mental states for those strict behavioralsits out there.

How would you define "mental states"? You can't have a mental state unless you have a mind, and I do think "minds" are limited to living beings. You wouldn't say a computer has a "mind." Certainly, are processing hardware might be quite replicable by technology, but the word refers strictly to living beings' processing hardware, not computers'.

[doogly]

Getting away from the pseudo-science, common usage is, again, a very bad way to define anything. In these debates, I tend to argue consciousness as the ability to communicate emotions, or perhaps the ability to communicate mental states for those strict behavioralsits out there.

How would you define "mental states"? You can't have a mental state unless you have a mind, and I do think "minds" are limited to living beings. You wouldn't say a computer has a "mind." Certainly, are processing hardware might be quite replicable by technology, but the word refers strictly to living beings' processing hardware, not computers'.

You can't a priori restrict minds and consciousness to biologicals, because that is what this discussion is about. You will have to come up with some argument as to why this is justified, not only at present, but imagining that one might encounter a computer intelligence capable of functioning equivalently to a biological one. If you exclude them by definition, that just shows you are likely to be using a too limited definition.

[Azrael]

This isn't (and never was) very philosophical. This is just Serious Business.

Tags removed.

[Probably]

Wow. So I came into this forum looking for information I could use on a project of mine.
This topic convinced me to register. But, anyways...

The discussion goes everywhere!

Correct me please if I miss anything, but it looks like the main problem is that consciousness is hard to define in a manner such that many people with varying viewpoints find the definition satisfactory. It's really easy to accidentally (or intentionally!) argue against a straw man and completely miss each other's points. That said, I like the decency that everyone's shown. It's very nice.

In order to sidestep the issue of defining consciousness, and to try to discuss the topic to a useful/informative extent, it looks like several people are trying to speak generally, which results in vague positions. I guess this is the way to go, though, pending future insight.

My own opinion (oh, no, another opinion!) is that there is nothing special about "human" consciousness and a proposed "simulated" consciousness. For some original (original to this topic, anyhow) content, I think that if we can interact meaningfully and gain mentally or emotionally from an experience with the "simulated consciouness" in question, we could call this... program(?) conscious.

On the other side, on a technical point, if we found everything to be indistinguishable, but at some point the program did something that we would call "definitely not conscious" (whatever that means), I guess it wouldn't be conscious; it's like the idea that a solid counterexample to a scientific theory forces people to re-evaluate and recosider the conditions of the explanation. The problem is whether we should call the program "conscious" pending an infinitely long decision procedure that tells us "no, the program is not conscious, definitely" or "so far, the program looks like it is conscious". I'd say that that is perhaps a moot point; (valiantly trying not to slip into solipsism here) if we can't even decide if our fellow humans are conscious, we can just call them conscious as a de facto state, and a similar reasoning would apply to the hypothetical program.


Do I make any sense?

[hideki101]

Here's another way to look at it. Say you created a self-replicating nanobot with the ability to accept and release chemicals and let it loose in someones brain. Once there, a single nanobot will "observe" a brain cell down to the atomic level, looking at all the chemical inputs and outputs and store the data to memory. After which, it destroys the cell and takes its place, sending and receiving signals exactly how the brain cell would have done. Now say, whenever two nanobots would connect, they do it digitally, rather than chemically. This continues until the entire brain (and possibly nervous system) is converted.

Would there be a point at which, the person stops being human? If so where? If not, what if this process continued with every other cell in the body (assume flexibility or rigidness as specific to an organ)?

My answer to this is that there wouldn't be a point where the person would stop being human. The thing defined as consciousness is the result of uncountably many interactions with whatever is nearby. As I said in another thread, in effect, any system of particles can be considered conscious to some degree. If you create a simulation of a brain to an arbitrary level of accuracy, such that it is indistinguishable from the workings of any other human brain, then it should be treated as a living human brain.

[Telchar]

I wish people would stop talking about rediculous limits like "Indistinguishable from humans" because they don't mean anything. If an AI is indistinguishable from a human, we wouldn't know so it wouldn't matter. The point is that we probably will not get to that point so in the mean time we have to address what we do with things that mimic intelligence without being indistinguishable.

Also,

As I said in another thread, in effect, any system of particles can be considered conscious to some degree.

No. Concsiousness doesn't have degrees. One either is, or isn't. Degrees of intelligence, a byproduct of consciousness, yes.

[0.0]

I would like to redefine the consciousness we are discussing so maybe we can get a consensus on the definition, otherwise we will be arguing about different things. Can we all agree that consciousness is the 'knowing that we are able to think?' Sort of an interpretation of the self awareness argument but seems a little more descriptive. I guy who translates chinese knows he is translating chinese, a room that translates chinese, whether there is a person running the system or not, does not know it knows chinese.

So if a computer can perfectly resemble me in a conversation but doesn't know it, it does not have consciousness. If it does know it, than it does.

If we can agree on that, then the question becomes, how do you know if it knows?

[Indon]

I wish people would stop talking about rediculous limits like "Indistinguishable from humans" because they don't mean anything. If an AI is indistinguishable from a human, we wouldn't know so it wouldn't matter.

The person who made it would obviously know it is an artificial intelligence - even if nobody would believe him.

The point is that we probably will not get to that point so in the mean time we have to address what we do with things that mimic intelligence without being indistinguishable.

That makes the discussion trivial - if it can't pass a test to demonstrate consciousness, then it's not conscious. If it passes all the tests for consciousness (as a human would, as I imagine we are assuming humans to be conscious), then it is conscious - and by definition indistinguishable from a human in this respect.

As I said in another thread, in effect, any system of particles can be considered conscious to some degree.

No. Concsiousness doesn't have degrees. One either is, or isn't. Degrees of intelligence, a byproduct of consciousness, yes.

Is a sperm conscious? Fetus? Newborn? One-year-old? Five-year-old?

If consciousness is binary, then we should be able to identify when it happens in a human.

Can we all agree that consciousness is the 'knowing that we are able to think?' Sort of an interpretation of the self awareness argument but seems a little more descriptive.

It seems more descriptive, but as it is qualia, it's meaningless to communicate. I can't even "know" if you know, let alone the chinese room. As it is impossible to percieve the qualia of other objects, we can make no meaningful judgements about the possession or lack of any qualia from anything.

So not only can I not conclude that you are conscious (as I can not percieve your knowingness or lack thereof), but I can't conclude that the Chinese Room is not (as I can not percieve its' knowingness or lack thereof) - it is as inscrutable as any agent, or indeed, non-agent.

All our judgements about if something is conscious or not must be based on our external observations of it.

[eMesreveR]

Can't we just say that our sense of consciousness is merely an instictive illusion to help ourselves process the world, like a soul, or a supernatural ruling being? As such, there's nothing inherent about it, it's just an artifact of increasing complexity and processing power?

Does that solve any problems?

[0.0]

I wish people would stop talking about rediculous limits like "Indistinguishable from humans" because they don't mean anything. If an AI is indistinguishable from a human, we wouldn't know so it wouldn't matter.

The person who made it would obviously know it is an artificial intelligence - even if nobody would believe him.

The point is that we probably will not get to that point so in the mean time we have to address what we do with things that mimic intelligence without being indistinguishable.

That makes the discussion trivial - if it can't pass a test to demonstrate consciousness, then it's not conscious. If it passes all the tests for consciousness (as a human would, as I imagine we are assuming humans to be conscious), then it is conscious - and by definition indistinguishable from a human in this respect.

As I said in another thread, in effect, any system of particles can be considered conscious to some degree.

No. Concsiousness doesn't have degrees. One either is, or isn't. Degrees of intelligence, a byproduct of consciousness, yes.

Is a sperm conscious? Fetus? Newborn? One-year-old? Five-year-old?

If consciousness is binary, then we should be able to identify when it happens in a human.

Can we all agree that consciousness is the 'knowing that we are able to think?' Sort of an interpretation of the self awareness argument but seems a little more descriptive.

It seems more descriptive, but as it is qualia, it's meaningless to communicate. I can't even "know" if you know, let alone the chinese room. As it is impossible to percieve the qualia of other objects, we can make no meaningful judgements about the possession or lack of any qualia from anything.

So not only can I not conclude that you are conscious (as I can not percieve your knowingness or lack thereof), but I can't conclude that the Chinese Room is not (as I can not percieve its' knowingness or lack thereof) - it is as inscrutable as any agent, or indeed, non-agent.

All our judgements about if something is conscious or not must be based on our external observations of it.

Your arguments seem flawed on the basis that we know as much as we will in the future. Your argument about fetus/5 year old for example: why does your list of ages prove anything? I am either missing the point or you are just assuming that since we don't know when consciousness occurs, it must be analog. Of course in the future we may know the exact moment consciousness begins.

[SpazzyMcGee]

Your arguments seem flawed on the basis that we know as much as we will in the future. Your argument about fetus/5 year old for example: why does your list of ages prove anything? I am either missing the point or you are just assuming that since we don't know when consciousness occurs, it must be analog. Of course in the future we may know the exact moment consciousness begins.

I thought it Telchar's fetal development example was relevant and very good. There are stages of consciousness. If you say one must be this smart to be considered conscious then where is that point as a human develops in the womb? A baby is conscious right? Well what about a week before it was born? A month? 8 and a half months? The same goes for evolutionary development? Humans are conscious. Dolphins, great apes, elephants and dogs are conscious. Are insects conscious? Are amoeba? Plants?

And what do you mean by what we will know in the future? There isn't some flipped on switch in our heads that says consciousness. You can't just say "we learn more and more every day, so one day we will find that switch". Saying that we are constantly learning how the world around us works supports your point is invalid. You can say that about any point and if it can support any point it supports no points. Why do I believe invisible pink unicorns live on the dark side of the Moon? Well we might one day find them... What possible theoretical method could we use to deduce the whether the "consciousness switch" is turned on or off in a creature?