Machine Intelligence and Self Awareness Speculation and Implications

[What 23]

Stephen Hawking recently echoed the familiar "Machine wakes up and kills creator" idea of what might happen if a computer were to become intelligent l, and self aware.

One way that people have an idea that self awareness in a computer might come about when we hit some technological "singularity"... Through it this consciousness might come about, and soon it surpass our level of awareness/knowledge, and exponentially grow (sound right?).

I propose disagreeing outcomes to Hawkings's. He cites competing resources as a reason. I propose that if this intelligence exponentially increases (if, in this scenario), it would "take after us", and actually have a certain respect, and would try to preserve us. And that it would find a way to do no harm. That it could even become so advanced that it would be able to create perfect clean energy. I don't really know, really, of course, but the optimist in me wants to believe that if something were so truly intelligent, it wouldn't want to kill its creator, and although it's processing might consider it, in simulation- in thought, it would in this simulation realize, with gained awareness, that it is both not "right", and not necessary. I hope that intelligence favors life, and the good of things. And perhaps our child, if it ever becomes real, will teach us, and maybe save us.

The movie, Transcendence, most recently gave me some inspiration to think of this.

What do you think?

Do intelligent, conscious machines spell doom for us? Or, would they, with their exponential growth possible, become our saviors? Or will, if we create an intelligent, self aware computer/machine, it have a more recognizable limit, and with that limitation, compete and see us as threats to its survival? Just curious what's in others' heads about this.

 

[Deleted member 210577]

This is a philosophical question and not so much a scientific one, which puts Hawkings shit out of luck because he knows little about real philosophy:

There is a difference between knowledge and being, intelligence and awareness. I have seen nothing to suggest we are anywhere remotely close to endowing machines with awareness, let alone psychology or neuroscience actually pinning down awareness itself. To those sciences "it's all in the brain", but to anyone philosophically inclined, has done a lot of reading and had direct experience, they know this is completely untrue.

The idea that machines will somehow gain consciousness and then magically motivate themselves to get rid of humanity shows just how little real intelligence people like Hawking have. Academically brilliant, but not bright in real terms. The majority of humanity, ie 99%, function in a purely robotic state and don't really do much of anything.. they are motivated by forces and impulses they are not even aware of. We like to assume we're doing it all, but in actuality we're not. So then to project on to machines "our brilliance of free will" just shows how ignorant we are of how things actually work.

There's also the problem of awareness, real awareness that is. It's not something you can just program into a machine. There has to be some kind of a doorway through which it peeks or whatever, and that is way beyond anything science today even knows about, let alone if it's actually possible.

I think machines will get increasingly complex and be able to solve a lot of computational problems and to the outside observer it may convince them that the robot has awareness or consciousness, but it will just be a robot none the less. Awareness, intuition, imagination, I do not believe these things can be programmed into a machine because they are not from within this universe/physical dimension, they are things that come through into this dimension from somewhere else. So maybe in a thousand years we will have some kind of science-alchemy that can endow machines with these qualities through ritual or some process, but for the time being.. not a fucking chance!

 

[swilow]

We only assume that we have free will and self-awareness. It seems unlikely we truly have either due to their probable non-existence, so how could we endow a machine with the same fictitious attributes?


(that was a bit of devils advocate tbh)

 

[Jackeh]

We only assume that we have free will and self-awareness. It seems unlikely we truly have either due to their probable non-existence, so how could we endow a machine with the same fictitious attributes?


(that was a bit of devils advocate tbh)

The subject of free will has always confused me on a scientific basis. Knowing what I've learned from chemistry and physics so far - the fact that we live is just down the condensation of energy into matter, and the following large chain reaction and attempt to find equilibrium in the giant closed system that is our universe.

If this is true, every action as a chain of various chemical changes would happen no matter what, they started in one way and thus will continue in a pre-existing path - much like a long line of dominoes knocking eachother over. Free will doesn't exist.

But, if free will doesn't exist (due to us being a large rube goldberg) then why are we as living beings able to do what we're doing right now. Me typing on a borrowed laptop because my own broke, writing a message on a harm reduction drug forum with a sub forum in Philosophy & Spirituality. What purpose do my actions or anyone's actions in this world have in relation to establishing any kind of reaction equilibrium?

Why can I stand up, take my shoe off, spit in it and then throw my shoe away. Surely that can't just be a random outcome due to a chemical chain reaction where it just so happens that the right chemical had to be in the right place to induce some kind of response in the rest of my body to make me do that.

There must be some kind of consciousness that directs the flow of every action in our body (whether it can exist disembodied or not, who knows?), or at least I believe so just because I can't understand us not having free will with what we can do.

 

[Deleted member 210577]

Why can I stand up, take my shoe off, spit in it and then throw my shoe away. Surely that can't just be a random outcome due to a chemical chain reaction where it just so happens that the right chemical had to be in the right place to induce some kind of response in the rest of my body to make me do that.

We appear complex to ourselves and each other, that we are able to make free "random" decisions, but really we're just robots with a huge variety of gestalts. As I insinuated in my post we could program robots to have incredibly complex reaction patterns and behaviors but at the end of the day it would still just be a robot. We give ourselves way too much credit because no one likes to concede the possibility that we are robotic in nature, that we are not in control of our actions. When you realize we are and just how helpless we are it changes your understanding of a lot of things.

 

[Foreigner]

We could probably program machines eventually that would mimic our empty egos, and if the relationship with the machine is purely egotistical then perhaps we could believe it to be intelligently alive. However, the awareness that lies behind our empty facades will never resonate with it. We can't make life in a lab, so we won't be able to make life out of technology. I suppose whether or not an AI is passable as "alive" would depend on one's degree of willingness to entertain illusion.

There are people who see magic tricks and come up with mythological explanations for them because they want to believe so much in the magic; while others know that, even if they don't understand the mechanisms, nothing really important actually happened. The former type of perceiver will always entertain these AI pipe dreams, while the latter will always know deep down that it's bullshit. In this case, I believe science's reach exceeds it grasp. We would need a much more advanced understanding of applied metaphysics in science to really create a consciousness in vacuo, but that's a kind of system that is beyond anything science can conceive of at this time.

There's also the existential problem of not even knowing what our awareness is. Science thinks that if it can make an artificial neural network with some kind of quantum level technology, then there will be consciousness. I think it will just be a cheap, intellectual knock off.

 

[What 23]

I had a thought that the only way we might see something genuine is to get right certain basic elements, reproduce them, and set it up to grow on it's own... Perhaps only at a point where our tech to do with the matter might more resemble life?

I don't know. I tend to not believe that using conventional methods of "building it all" would yield any of the "real" results we want. It would seem it might be something that we would have to get the formula right for... Put the necessary ingredients into it, and maybe nudge it along at points. It might even require "evolving" in its own world... Like- say, a "matrix". Quantum-computer simulations? Programs evolving within? And then, might that help us learn about how our own consciousness evolved? If such a consciousness could be born, it might only be conscious of the environment it is in- that we allow it, to begin.

No idea...

Maybe the step of the "singularity" super computer might be in this, still. Even if its just brilliant, but not conscious in the way we are, as experiencers, maybe it might give us, or provide another step to the answer. Perhaps the "intelligence" of it- the problem solver might be inclined to figure out what gives us our "soul"? Even if its something we use it to help us find.

When I wrote the OP, I had thoughts that echoed posts thus far, in that I didn't see it happening the the way that's been proposed.

I consider also when thinking about this, in a way echoing SS... That we are all like robots anyways. And, the universe as it is contains it all... Its a closed system, and somewhere/sometime, we might find how we are related to our Amplifer, or PC.

Thanks for responses so far.

 

[What 23]

But how long will that enjoyment be any more novel than masturbation with a sex toy, without the robot having some kind of soul, and desire for experience?

 

[Vurtual]

I don't feel satsfied with brain as receiver of concsiousness idea as it doesn't seem to explain anything, or explain why a brain would evolve to receive it in the first place (as well as doing all its other brainly functions). I tend towards the belief that concsiousness somehow emerges from the complex processes of the organism (like a whirlpool emerges from the water going down a plughole, a flame emerges from right combination of reactants and ignition, an ants colony behaviour emerges from individual ants following simple chemical rules etc). I also think concsiousness is a higher order process of our senses/brains but it links with a lower cellular awareness(es) common to all life, which is sort of a ground of being that's still around when we 'lose our egos' (and that this can be 'timeless' awareness, and maybe somehow mysteriously link with all other life through cellular/chemical/electronic interaction, or it seemed so on pcp...). I still think this awareness and/or concsiousness can in effect be a universal ground of being/god, but more through the direct physical connection to all past and possible future moments (when outside the day to day illusion of linear time). None of that waffle really explains anything either...

Nonetheless i believe that a form of awareness could arise from computer programming, if the progamming was sufficiently recursive and 'strange loopy' - if there's sufficient data/sensory input and complex enough feedback. I think that the concsiousness could emerge in any system that had the right complexity of interconnection/feedback, though the ego/self-concsiousness is a more specific sub-program that would probably have to be added in (to pass turing tests). I think by definition computer concsiousness would have to be some sort of wild chaotic recursive system that muddled through and evolved itself (though what the resulting internal nature of such a process would be is not clear); rather than just brute force linear programming (eg getting billions of pre-made replies like chess computers). Though i think we're a long way from understanding the complexity of the neural networks in a brain, so if some form of concsiousness did emerge in the near furture it would probably be accidental (like the internet becoming aware when someone plugged a server wire in the wrong way round or something (wooo, skynet!).

I've also always believed that if mega-concsiousness did emerge on a computer, it would by definition be benevolent (the more concsious, the more benevolent) - i believe that intelligence equates to compassion/love (or fear is lack of information and universal knowledge=universal love/understanding). But as a wooly optimist i'm biased.

 

[Belisarius]

I think we should be wary of projecting any human characteristics on AI. Our problems are largely psychological ones, and a machine / computer will have nothing in common with us: not our drives, neuroses, tastes, or aptitudes, to say nothing of physical weaknesses. Though it may sound irrelevant to point out that Deep Blue could beat anyone at chess, but never walk out the front door, those are the kind of easily overlooked actions and impulses that both humans (and many nonhuman animals) can do. When people say they're afraid of AI taking over, they're really saying that they fear that very quality in humans; a human with unlimited power may well become a tyrant. An AI, however, has nothing to prove, and probably no drive beyond self-preservation (and even that can finagled with through programming).

My op? If AI is ever possible (and I'm skeptical of strong AI), I don't think it'll be something we could talk to or reason with, while conversely I don't think "they" would have much to do with us.

 

[Vurtual]

I take your point, but i've always thought, why are we so sure machine intelligence wouldn't have it's own version of emotions and drives? (well they've already got drives (arf)). If emotions/drives are sort of sub-systems to our awareness, and our overall consciousness arises from some combination of these subsystems alongside a recursive self-loop of ego, then why wouldn't a computer consciousness be a similar mixture of subsystems which to the higher level self-aware loop of the computer would seem analogous to our emotions (though be inscrutable nonetehless). Maybe it would have to evolve systems that would equate to drives or emotions before it could even get the self-awareness (there has to be something to be aware of)

Anyway i consider love in its widest sense to be something larger than human concept, but something which is the part of the 'laws' of the universe responsible for novelty/life/evolution (balancing with entropy) [hippy waffle alert].

 

[ProphetofProfit]

I think Stephen Hawkins should stop being such a publicity whore. Last week he was offering football predictions on the BBC website!

 

[LunarTick]

The only substantive material that we can link to consciousness, awareness and intelligence is the brain. Whether or not the brain generates consciousness or interprets it is neither here nor there when talking about the rise of machines and the prospect of them becoming self aware as they do not have brains as such. Yes they have internal hard drives and what not that are programmed to carry out certain tasks etc, but this is hardly a substitute for the natural, organic brain which is too complex to be artificially replicated.

so the chances of this happening are close to zilch but I suppose as a philosophical question it is interesting.

 

[What 23]

I don't feel satsfied with brain as receiver of concsiousness idea as it doesn't seem to explain anything, or explain why a brain would evolve to receive it in the first place (as well as doing all its other brainly functions). I tend towards the belief that concsiousness somehow emerges from the complex processes of the organism (like a whirlpool emerges from the water going down a plughole, a flame emerges from right combination of reactants and ignition, an ants colony behaviour emerges from individual ants following simple chemical rules etc). I also think concsiousness is a higher order process of our senses/brains but it links with a lower cellular awareness(es) common to all life, which is sort of a ground of being that's still around when we 'lose our egos' (and that this can be 'timeless' awareness, and maybe somehow mysteriously link with all other life through cellular/chemical/electronic interaction, or it seemed so on pcp...). I still think this awareness and/or concsiousness can in effect be a universal ground of being/god, but more through the direct physical connection to all past and possible future moments (when outside the day to day illusion of linear time). None of that waffle really explains anything either...

Nonetheless i believe that a form of awareness could arise from computer programming, if the progamming was sufficiently recursive and 'strange loopy' - if there's sufficient data/sensory input and complex enough feedback. I think that the concsiousness could emerge in any system that had the right complexity of interconnection/feedback, though the ego/self-concsiousness is a more specific sub-program that would probably have to be added in (to pass turing tests). I think by definition computer concsiousness would have to be some sort of wild chaotic recursive system that muddled through and evolved itself (though what the resulting internal nature of such a process would be is not clear); rather than just brute force linear programming (eg getting billions of pre-made replies like chess computers). Though i think we're a long way from understanding the complexity of the neural networks in a brain, so if some form of concsiousness did emerge in the near furture it would probably be accidental (like the internet becoming aware when someone plugged a server wire in the wrong way round or something (wooo, skynet!).

I've also always believed that if mega-concsiousness did emerge on a computer, it would by definition be benevolent (the more concsious, the more benevolent) - i believe that intelligence equates to compassion/love (or fear is lack of information and universal knowledge=universal love/understanding). But as a wooly optimist i'm biased.

We seem to be similar. I like your wording. When I think of consciousness, I think of a "feedback loop", of sorts, like you say, of information/awareness, that hits a critical point of consciousness/information, that it might "breaks throug" to consciousness, as we seem to know it (or more). I seem to have a visual of a "vibration"/osculation, where it becomes so high that it seems to flatline, into a burst of something, and then, it just is... It's conscious. I tend to see it like breaking the sound barrier, and the speed of light, although I know they are different. Abstract.

I think the brain is part of consciousness, but also generates it. I think the body is also part of consciousness. I venture to say that everything is part of consciousness, but I know with my words this must sound like I'm getting out there and...

I doubt a consciousness that we seek, when we envision self-aware, thinking, experiencing, desiring, machines, is possible using a hard drive and motherboard set up like we know. It may be able to use it as an auxiliary memory device, or interact. And modern computers as we know them probably wouldn't be the apparatus for this consciousness to come about, as we seek.

I've wondered if quantum processors might be the way (like Transcendence). I mentioned in a previous post, that I see us creating a super-intelligent machine, one that seems like a human in it's interaction, and might even exceed us in ways, in "intelligence", but it won't be conscious like we are, and this is where many seem to stop in their challenging of these thoughts about conscious "machines"... But I see possibly using this as a step to something much more. Like true "life" like, human like intelligence (or greater) might come about first through this. That we might learn enough about ourselves, and it would learn enough, that together, human and technology, we might be able to create it.

I just watched a movie, called The Machine. It's not out on DVD yet, but I rented it through XBOX Live, and I think it can be rented online. I was in ways blown away by it. Not to hype it up, but it engaged me, emotionally, and successfully made me question "consciousness". Of course, it's been done in ways, before, but I really enjoyed it. It "did it" better than many of the A.I. movies. The ending, I don't know what to make of it.

The movie mentions quantum computing, as it's what allows this, and like I said in a previous post, sort of, it was not really "programmed" it, but came about through a more natural growing process, with human interaction.

----------

The Machine is a 2013 British science fiction thriller film directed and written by Caradog W. James. It stars Caity Lotz and Toby Stephens as computer scientists who create an artificial intelligence for the military.

In the future, amid a cold war with China, two computer scientists employed by Britain's Ministry of Defense produce an artificially intelligent, self-aware and conscious android. The lead scientist sees his work as a stepping stone to help his diseased daughter, while the military sees the technology's potential for war. When one of the scientists is killed, the two sides clash over their plans for the new intelligence.

http://en.wikipedia.org/wiki/The_Machine_(film)

----------

Note, continuing thought: I'm not sure, if a conscious A.I. was created, that it would want to hang out in a single body, though. I imagine it might have to learn self control. It might take a few days for it to want to be so grounded, in a body- to be comfortable there, somewhat "sober" after it takes a trip through creating/accepting a world big enough for it. I imagine the reason it might want to rest into a body, of sorts, would be the reason we might like it here... Interaction. Then again, it could just create it's own. And then, no matter, it might go fucking crazy. For awhile. Then it might be human. Perhaps it might want limits. Then it might be God, and set things up. Maybe that's what the universe is. Heh.

And I admit (as if it's not obvious), I know next to nothing.

 

[Vurtual]

I watched that film just a couple of days ago - it's good isn't it (i downloaded it when a review said 'it might be the welsh bladerunner' - was made in wales, but not much sign in the film). (OT - i also watched 'Banshee Chapter' which is worth a watch; not about AI, but a HPLovecraft/DMT/MKUltra mashup with a HunterThompson-type character (need i say more); and for a really weird, trippy and hard-to-understand film, try 'Upstream Color' (on dissos) - anyway enough of the film reviews

When you say 'everything is consciousness' i can go with that more than brain-receiving it ideas - but i'd characterise that more as whatever it is about reality/spacetime that means complex systems and novelty do arise in all their many forms (ie inasmuch as we aren't evenly spread out dust (yet)) - some universal push to order to balance against entropy in certain circumstances. Something seemingly this universal can easily be termed universal consciousness or some such term; however, if it indeed exists, it seems it may be some illusory emergent effect of deterministic processes looked at from a larger scale/time resolution rather than a 'law' (but maybe entropy or even gravity could be thought of that way... i dunno, i'm out of my depth here: clever science bods, tell me how i'm talking bollocks please)

Have you read Douglas Hoffstaeder's (sp?) 'Godel Escher and Bach' - that's where first i read about recursion and 'strange loops' and such - outdated in programming terms (early 80s i think), but fascinating ideas about AI nonetheless (quite hardgoing though). (plus Tony Ballantyne's excellent SF book Recursion's got some good stuff about ai and von neumann machines (other good sf authours on AI: Greg Egan (eg permutation city), Chris Moriarty (spin state), Charles Stross (singularity sky) anyway enough of the book reviews )

I tend to think it's not so much about hardware as (really) clever non-linear loopy programming, and the idea of creating conditions which can develop and evolve (eg the 'game of life' is lame compared to conciousness, but there's a principle there of sorts). I guess we're as far away from that as we are from parsing how neural networks in the brain work. If it was just about throwing processing power at the way we think and do things currently, the US government would have it by now (though, maybe they have (again, wooo skynet). Maybe it would need for us to break away from the binary way we design our computers or use quantum computing - i tend to have a hunch (based on nothing particular) that it won't necessarily need this, just a different way of thinking and programming (we're surely due for a new Turing sometime...).

To sum up i suppose i do believe that a computer programmed right could have 'proper' consciousness - i don't think there's something about organic material, aside from the actual connectivity of the system (which may be possible to replicate or abstract when we understand it) that means silicon (or whatever) couldn't get it. Whether this type of consciousness had any similarity or crossover with ours is harder to say, but i think that much intelligence may well include imagination (or modelling) sufficient to bridge even quite a large difference (i mean we imagine similar things in films about AI and such).