Can AI eventually become 'self-aware'?

First, the question posed can an AI become self aware. Sentience isn’t necessarily required, of if it s by some objective measure, then make a logical argument for that. You still haven’t supported a logical argument for the requirement of life as a necessary for self awareness, but have passed that discussion to making the assumption everyone agrees it is when you categorically reject that an AI can become self aware. You are very far from making that argument.
You use a lot of words for which you make undefined assumptions that others will share your meaning without a discussion of definitions. Let’s go that road, see how you support it for self awareness, sentence, and even consciousness. I find it interesting that you make declarations of meaning and regarding concepts that are still massively debated in academia, science and philosophy. For example, there is a growing community of panpsychists that do not believe biological life is a necessary criteria in the discussion of consciousness, a view with followers publishing in journals like Scientific American and even some advocating their positions on YouTube with far better persuasive logic based arguments than you’ve made like John hagelin, Roger Penrose, and Leonard Susskind.
Without establishing a consensus on objective definitions and the concepts of consciousness, self awareness and even intelligence, we remain embroiled in arguing over the number of angels that can stand on a pin.
In regard to Sophia, the capabilities she exhibits forces us to more closely examine what we mean by the concepts of self awareness, consciousness, and intelligence. How would we measure if something possesses those attributes?
Is Sophia self aware? Depending on semantics a the definitions of some, it would indicate she is, yet because of the ambiguity involved in defining those abstract concepts others would say no. That a consensus doesn’t exist yet, if there is one major lesson regarding Sophia is she is forcing us to think deeper.
As for Sophia herself, here’s what she has to say. I find her response interesting.... with thinking about.
[video][/video]

 

I would define sentience that which is endowed by nature with life, that which is alive.

How is that subjective? Can you not sense the difference between life and a machine simulation of it?

All life is biological. But, where I differ with science is my belief that life has a spiritual basis. No way to falsify that, though.

But, since my understanding of life has that orientation, I know realize that I shouldn't have started the thread (or at least, not in the science forum), because it becomes a question of spirituality, not science. But, since my belief is not rooted in dogma, I'm not seeing any conflict with science on the point, other than the idea that scientists believe AI 'self-awareness' whereupon a human cannot tell the difference between an AI and a human, is therefore 'alive', I reject that concept entirely as that is not proof, and that it is not proof is logical. But, scientists believe that self-awareness by a machine is possible, and I declare that it is not, at least, not without aid of a substantial quantity of organic matter controlling the system (a cyborg, as it were, but which puts it back in the nature column as the 'self aware' arises ultimately from the biological, not the mechanical).

I am not an 'authority' for anyone, but I do rely my own spiritual experience and perception for my understanding.

 

It doesn't matter how 'interesting' i.e., complex, sophisticated, Sophia's response is.

Let us all live together for a few days, and I gaurantee it will become evident that her machine limitations will become evident.

Why? Because the programmers cannot think of everything. It's the same problem Uber and Waymo have with their AI self-driving cars. They manage the terrain, self learn, but there is always something the programmers didn't think of. In fact, that which programmers do not think of, though they think of a lot, the machine self learns a lot, but that which they do not think of is, in terms of potential number of variables, is infinite.

See? A human's response is infinite because there is a sentient being driving the organic matter. It is my belief, however, that that sentience has a spiritual basis. But, since science will not recognize that fact, it will be forever chasing it's tail, so to speak.

There is no such thing driving the machine, other than the cleverness of programmers and since that is finite (or rather, finite at any given moment, noting that programmers will continue to make the program more complex, so we have a case of the finite chasing the infinite, and it will never reach infinity, it is impossible, which goes back to my original assertion, that the distance between a machine and life is infinity because life response is infinite and will driven, and machine response is finite, and man is the intelligent designer, so we have a case of man trying to compete with nature. It's not a contest. See? Nature has been around for what, some 14 billion years? How can man even begin to think he can match that? I don't see it, personally), and so living with a machine will eventually reveal limitations programmers did not think of. See? It comes back to programming, which is done by humans trying to achieve what they don't understand is impossible, machine equivalents of life itself, which is impossible in a total machine system. At the very minimum, a cyborg creation will be required, but that puts it back into organics, whereupon life is endowed by nature, not man. Man can tamper with and engineer genetics, but man isn't the creator of life, nature is, and that is that ( noting that I do not believe in 'intelligent design of life, i.e., 'god').

It all comes back to the human.

 

You are correct with your earlier post. While the topic could be an interesting discussion in a science forum, you have approach the subject by arguing from a ‘spiritual’ subjective perspective summed by the statement, “I don't see it, personally”. Though you indicate you aren’t arguing intelligence design, we might as well be arguing about God, given there isn’t a means to bridge your personal, your subjective feelings from a science based perspective. Just as there is no common base for a cross epistemological framework coherent analysis between that the uses the scientific method to understand reality and that is used to discuss the supernatural, there is no basis to discuss your objections to the possibility that An AI can become self aware. In your paradigm, only a biological can possess self awareness by definition, so end of discussion. Among those that explore reality using the scientific method, one of the first things that must be defined is what consciousness is and if there is a point where all the criteria exist for consciousness to occur, for instance, is consciousness in humans emergent from a certain threshold of neural connectivity or some other combination of factors.
Yes, billions of years of natural selection might be involved through natural processes, but that doesn’t mean billions of years are required to understand the structure and functionality that was the end product. However, scientific study is piecing together how neurons work, communicate, and how they are organized and how they are maintained. The architecture is being replicated, currently being tested and demonstrated, and some of the existing designs work, not because programmer coded for every possibility, but because they replicated nature’s design framework, including the ability to learn as does our neural frame work. The neuralmorphfc chips, designed using the currently understood architecture of a brain’s neural frame work represents a fraction of a human’s trillions of synaptic connections currently the capacity of a species of worm, but the design is scaleable, and with the advent of combining quantum technology with neuralmorphfc chip design, potentially able to operate a speeds similar to biological neurons. So, as the design is scaled, it become potentially available to be used to better understand how to improve the design, to more rapidly scale it, and to improve it’s overall capabilities, like designs that can receive sensory information (already being tested in robotics). So, the question is, can we learn how consciousness occurs, what is required, and whether it is an emergent thing and if so, at what threshold? The more we build, the more questions we have, but little by little those questions are being answered.
I suspect you don’t have a clue how far understanding of either neural science has progressed or how rapidly and to what level technology has progressed seeded by what has been learned.
But, to determine the possibility that AI can become self aware, develop consciousness does not mean it is alive in a biological sense, but will depend on refining what we mean, how we define those concepts. Buy some definitions, casual ones, Sophia is self aware; she know what she is, can describe how she works, she learns, and can even describe what happens when she is switched off. She operates autonomously and can seamlessly hold conversations with people on any topic. Some might say she is self aware, but, none would suggest she is alive in the biological sense of the word... note would she. And for whether she has consciousness, we can’t prove any human truly possesses consciousness except to assume others have the same experiences as each of us do. While we can have a sense of consciousness in ourselves the best we can do at this point, is use our experience with other humans as some ambiguous frame of reference in determining what other things might possess what we hazily call consciousness, yet even in that there isn’t a consensus. Thinking through that, presents the same dilemma as Descartes found when exploring the questions regarding himself and in trying to rule out the possibility of his thinking being influenced by a manipulative demon (think programmer).
So, the topic is far more complex than just your beliefs, your feelings... it is a debate thousands of years in the offing, but we don’t yet have the precise language nor the benefit of precise definitions for true understanding what is or is not possible; If we don’t have consensus on what is or isn’t consciousness, how could we?
People like John Hagelin would suggest, the universe and everything in it is a conscious entity (some use that argument to define the universe as the super intelligence they call God) and that consciousness is a property of the entire universe and consequently everything in it, suggesting like some in ‘primitive’ groups that everything is conscious, everything has a spirit. If so, Sophia is composed of parts that those ascribing to Hagelin’s ideas, that all possess consciousness.

Here’s a thought problem. Are you a single entity with a single consciousness or, consider you are comprised of billions of living cells, a colony of sentient, self aware, conscious life forms. If you lose a limb, are you still....you?

 

Self awareness is consciousness, it is life. No machine is alive, it's just simulated.

As for your last question, actually, we are both. At which stage of organic development is a sentient being synthesized, no one knows. Lower life forms are 'life' are not 'self-aware' but they are alive. Humans are the only organisms that are 'self aware'.
No machine is self aware, it only simulates it. Self-awareness begets sentience, humanness, capable of free will. Humans have an ego, machines do not, but if they do, it's only simulated. Note, 'self awareness' is not a prerequisite for 'life' nor is it a prerequisite to free will but life is a prerequisite for self-awareness and free will.

Hagelin sounds a lot like a pantheist, and that may be reality, but though I'm friendly to the idea, I just don't know. I've actually claimed to be a pantheist, in the past, that the universe, itself, is one big conscious being. But, I doubt it --- I'll leave the door cracked open on that idea.

As for losing a limb, not only am I still 'me', the 'me' can exist separate from the body nor is a body required for the individual to exist. The body is like a cocoon for the soul, and, that is, in fact, in my view, it's purpose. I've had out of body experiences, and that is why I believe life has a spiritual basis. You can chalk it all off to imagination, or a dream. But, should it ever happen to you, I guarantee you will sing a different tune.

 

Yep, pointless.

 

Like I said, AI is not about humanness. And how is the human body so different from a 'machine'? A machine is just mechanical power, but if you connect a computer then you have a 'mechanical system'. A human body is controlled not by a computer but by a brain...synonymous with a computer. So our brain does function like a computer in which it processes information that it receives from sensory input, then sends messages back to the body as necessary. A computer and a machine can do the same process. But some will say humans can think and experience emotion, etc. and to this I say give AI a couple hundred thousand years to evolve, as modern humans have, and you might be surprised...

 

They won't reproduce?? Evolution will come from 'learning' and 'adjusting'. It's not about someone 'allowing' something?? It's about giving AI the ability to learn and adjust and given enough time how will we know their limits? My example was creating AI soldiers who can kill...maybe if the enemy yells 'uncle' the AI machine will retreat?

 

Then you have no thread?? AI is not a human...period...

 

Not according to this essay by Robert Epstein ( https://aeon.co/users/robert-epstein )
https://aeon.co/essays/your-brain-does-not-process-information-and-it-is-not-a-computer

I contend that the distance between a machine ( without organic interfacing, i.e, 'cyborg' etc ) and 'life' is infinity, i.e., Moore's law will not get you there -- There is no linear progression from one to the other, only simulation. It's not about 'what a machine can do' versus 'what a human can do' it's about what constitutes life, that is it a quotient beyond the reach of machinery.

 

I guess you will need to define 'life'?

 

It's kinda like porno, either you know what it is, or you don't, but don't ask me to define it.

 

Yep, pointless.

Opinions aren’t science, it they were, I’d be emperor of the world... no, far more than that, and you but a illusion, which you are... prove me wrong.

 

Prove me wrong, if and when you ( mankind ) create a machine that is 'alive'.

But, first, you need to learn what the difference between 'life' and 'machine' is. The former has life, the latter has no life.

The difference between you and me is that you cannot tell the difference, and I can.

The reason I assert that is that you believe it is possible for a machine to be created which is 'alive' or at least, no distinguishable difference.

But there will always be a difference, because something that is alive can respond to an infinite number of different stimuli in infinite ways, without programming.

A machine will always reach a finite limit, beyond which, it won't know what to do without more programming. Therefore, the machine, merely an extension of man, therefore is the revelation of man's limitations. This is not to say that some day, man will create a humanoid/android creature. But, that is similar to how man created corn from maize. That is genetic engineering, guiding nature to new forms. But man, in this instance, was not the one who endowed the organism with 'life', no, nature did that, man just took a steering wheel, nature supplied the vehicle.

In short, the machine is man attempting to create itself, but it will never achieve it because life cannot be acquired in a machine, because man doesn't understand the difference between life and the machine, so it will be like the proverbial dog chasing it's tail.

The difference is that life has a spiritual basis, it's home is in the infinite realm, whereas the machine is in the finite realm.

There is no linear progression from the former to the latter; it is beyond the reach of Moore's law. And no, science cannot explain it, nor will it ever be able to explain it.

To learn the difference, one must look to the sages, Gautama Siddhartha, Lao Tzu, Ramana Maharshi, etc., plus years of meditation, for those answers.

Nature will allow man to garner many of her secrets, but the nature of life, itself, is a mystery -- and I doubt she is going to give up that one, she's going to keep it for herself.

 

Not quite. Consciousness is only one dimension. There are clearly ways to detect where a brain fits in with the definition of life.

The issue here is limited to the consciousness part. Whether or not it can be determined to be alive is a very different question.

No, there is nothing infinite about our brains. Nothing. Our brains are meat machines lodged in a protective shell with various input and output devices for detecting light, touch, the vibrations of sound, etc., and for controling movement inside and outside our bodies.

I think you have a dated version of programming. Serious progress has been made in the programming realm in implemnting what is necessary to learn - not just to learn facts, but to learn strategy for responding to information. So for instance in the case of the game "go" the programming wasn't about how to win at go. No effort was put into that. It was about how to allow the machine to learn by experience in creating strategy. And, the result was that this game which seemed so impossible for a machine to address (FAR harder than chess) can beat every human champion. And, strategy elements used by the AI have been witnessed where humans couldn't even figure out why the strategy was winning.

Also, let's remember that human brains come with original programming, too. Our brains have the programming required to allow us to learn. We're born with about zip for knowledge. But, we have brains that have the physical machinery to do such things as build a dictionary and parse grammar. We have machinery that can figure out cause and effect, etc.

If you're really interested in how brains work, I'd strongly recommend books by Steven Pinker.

He has serious academic qualifications in cognitive science and relate fields.

In "How the Mind Works" he describes what we have as evidence of how our brains solve various problems - even including emotion. He states that he doesn't claim this as a "correct" definition - just that enough is becoming known that he believed it was time to compose a computational model that corresponds to current understanding.

His examples and descriptions draw heavily on real world medical cases. Obviously, one can't zap pieces of someone's brain and study whats now missing. But, there is a huge body of descriptions of deficits to draw on.

I don't believe he got everything right, but lookig at the brain from a computational point of view is seriously illuminating.

Besides everything else, let's remember that we got these brains from evolution. They probably started out as response to light in protozoans. Our brain was built as a step wise improvement in machinery.

 

The discussion you may having is in regard to between 'life' and 'machine' but I haven’t made such an argument. You equated life, biological life, as a fundamental requirement for developing self awareness, without any logical definitions to support that other than you have a ‘belief’. I did not make any suggestion that a matching could be considered life, but a couple times you have suggest I an others have... didn’t happen. Regardless, if I made no distinction, again you extend the argument to make suggest you know the difference, but offer no logic based discussion to show, beyond your subjective beliefs, that you do. Only your assumption without accompanying that self awareness requires biological life leads to any discussion of a distinction biological life vs biological life without any discussion of the requirements that define what constitutes life, an entirely different discussion than the OP asking if an AI can achieve self awareness. That failure of logic is analogous to the old logic failure of asserting all crows are black birds, therefore all black birds are crows. For most of your assertions you have provided the base definitions and arguments to support your position, just your ‘belief’. With out any agreement on the meaning of fundament terms and their definitions there can be no logical basis for discussion. That you don’t understand that I find interesting. Find another appropriate forum to discuss the concept of life; that should be interesting. Ask, a simple question, what is life? Think you ‘know’ the answer? Ok, fun... Google that question, see if there is complete consensus on that. If there is no agreement on the requirements that can be used to determine the concept of self awareness there is no means for assessing if an AI can exhibit or not, the attribute of self awareness. And, if biological life is required for self awareness, does that mean all life is self aware? There are those often speaking in an undefined supernatural sense that would say yes. Humans discuss reality with a common language and a consensus of meaning shared in the language, so to discuss the. OP, we haven’t set the proper basis for discussing the phenomena of self awareness enough to answer the question.

 

A bacteria fits your definition of life and yet it has NO "spiritual basis" nor does it have any "home" in the "infinite realm" which essentially places that life on the exact same level as the machine in your statement above.

 

If you defining 'consciousness' as merely being activated, 'switched on', then okay. But, I'm equating consciousness with self-awareness, and, as such, with 'life', i.e., that consciousness, by default, requires a living organism, which is to say, a machine can simulated consciousness, but it's not actually conscious, not in the sense of a living, sentient, being.

The "I" part, that part which is "you", is NOT the brain. Close your eyes, imagine a large elephant. What is looking at the elephant? That which is looking at the elephant, is YOU, and the 'you' is NOT the brain, you are NOT your body, you are something separable from it. And the only reason I believe this because of having had out of body experiences. Sure, it's not falsifiable, other than the fact that the dentist was startled when I asked her about what she was talking about when I was 'under' as she couldn't figure out how I knew what she was talking about. Well, I told her i was 'out of body' and remembered her conversation with her assistant, she was talking about some personal thing, an appointment she had to go on right when she finished with me.

I imagine science knows a lot about the brain. But the brain isn't the entire story of consciousness and humanity.

I, for one, do not buy into the premise that a human is nothing more than the body, and there is nothing beyond it. I believe in the human soul, and that it is eternal. Now, that's not the proper subject for science, nor can science even speak of it, as it's not containable by instrumentation, measurable, falsifiable, and all of that.

But, there are some other arguments against the notion that a newborn child's brain is an empty slate.

https://aeon.co/essays/your-brain-does-not-process-information-and-it-is-not-a-computer

 

I do not know at which point along the upward progression of the food chain, whereupon the soul in man is synthesized, or, rather, that point along the progression of advanced organisms, when nature endows an organism with spirit, but, wherever it is, ( and it just may very well, indeed, go down that far, who knows? ) but wherever it is, the 'infinite realm' is planted therefrom. However, I believe it does. I believe that the soul of man, like other living things, starts out as a small seed, and progresses, grows, until eventually it reaches the human level. This is the theory of reincarnation as I understand it, personally. Our souls are synthesized in small living organisms, millions, if not billions, of years ago, and we reincarnate upward, right along the evolution of organisms, until we are reincarnated into modern humans. In other words, as souls, we are very old. Nature, in her wisdom, erases memories of past lives. This is all not appropriate subject matter for science, of course, it cannot deal with such things, but it is my spirituality and belief.

 

When the rubber hits the road, my view on the matter is not the proper forum here in 'science'.

I shouldn't have, therefore, started the thread here. It should have been placed in the 'religion and philosophy' forum.

 

Exactly. This is really a religious question, since you've brought the soul into it.

I, and I assume most of the other participants in this thread assumed you were speaking in a purely scientific viewpoint, without spirits ghosts, and souls involved. You are really asking a very different question than what I assumed you were.

So in answer to that question, no, AI can't become self aware unless we figure out how to invent an artificial soul.

 

Ah... one of the questions somewhat explored in the movie, Blade Runner, and the book it was base on.
But, here is a tangent question regarding the concept of a soul, if a human is the result of being cloned, doesn’t it have a soul? And a follow on, can a cloned human, basically a man made product using specialized technology, be owned by the maker?

 

No, I'm talking about the full definition, including self awareness.

I just don't see how your "life" requirement applies.

Yes, our brains do strange things when deprived of senses (or when messages from our senses are scrambled by approaching death), when deprived of energy source, when parts of the brain are slowed or stopped with chemicals. It's not an "all or nothing" situation. We sleep, we dream, we can be rendered partially unconscious.

I think it is reasonable to dodge the failure modes when considering the central question. Whether an AI can dream or how an AI performs when its sensory inputs or power supply are failing.

Your article is right that babies come with programming built in.

But, he immediately makes the same blunder that he accuses the anients of making - that of imposing his own preconceived idea of how a brain "should" work without ANY effort to look at how a brain DOES work.

Modern brain science doesn't do that. The investigation being done is an examination of the brain to see how it carries out particular tasks. Of course, we have far more tools for doing that today due to huge advances in science based medicine. Some of that is really difficult. But, sometimes it's not so hard. For example, you may know that there are people with no short term memory or no long term memory. Sometimes peoople are born that way and sometimes it comes due to some sort of trauma. So the idea of there being one form of memory (like your author states) gets excluded, because we KNOW it doesn't work that way.

Another mistake he makes concerns how images are stored. Our brains actually separates "lion" in general and actual total renditions of a particular lion. And, that allows information about lions to be centralized. Then, if you want to remember a specific lion, the unique parts get stored in ways that those differences/improvements can be found.

But, almost alwyas a human isn't interested in a picture of a speciric lion. Just like a human is almost never interested in the actual picture of a dollar bill. Almost every important aspect of either of these objects does not come from a specific image. So even with images, the root image stored is generalized - think of an outline that then comes with various versions of more image information. Then, a human can take a glance and say "lion" or "snake" or "tree branch". The outlie is probably good enough to know what actions to take. Having pictures of those objects isn't really a help in the life of a human. It would only slow things down. When you use money you don't have to know what is written on the bill - except for some time you get asked for more image information. Even then, you almost never want the full image of a bill. The author's question about bills didn't differentiated between the images of the stiff new bills you gave your kid for Christmas or the billy you found in the guter when walking home. Almost always, you're interested in atributes, not images. So, even most of an image gets stored as atributes.

If you want to know how the brain works, you have to study the brain! Your author didn't do that. So, he's absolutely and totally wrong.

 

I resist moving this to religion on the grounds that in religion there is absolutely zero interest in how the brain works.

Any religion can state that there is a "soul" in humans regardless of any possible physical factor in humans or any other life form, or anything found concerning how the brain works. And, religion can say it is required for ... everything.

Religion always creates a catch-all solution. In this case, whatever isn't immediately understood and agreed can (and is) simply be assigned to the function of the "soul".

 

Why did you ask the title question?

Here I thought you were being serious!

Yet, it turns out that you don't even have a basic interest in how the BRAIN works, let alone what AI might be capable of supplying over significant time.

I'm not even slightly interested in dissuading you from your religion.

But, I will suggest that it's a little strange for you to make statements about what the brain can't possibly do when you aren't even interested in how the brain works.