Artifical Intelligence and Morality

The other day as I walked through the library I came by a book that sought to open the conversation about future AI intelligence. In particular, the current craze (y) with transferring our human intelligence to a super-computer. Certainly, for someone like Ray Kurzweil, this must happen. It's a fact of Moores law. Human beings are moving closer and closer to transcending our biological form; to forming ourselves online, in a computer generated reality. Easy-squeezy. I just hope - or rather, Ray Kurzweil MUST believe - that consciousness is nothing more than the exchange of information in a very large network.

However, there are some very basic logical and philosophical problems with transferring consciousness to a computer-substrate, and in particular, there is something fundamentally incongruous in conceiving "artificial intelligence" as 'intelligent' in a conscious sort of way.

Why do I say this? This is what I find amusing in todays formalized market-place of ideas. A technologist and innovator like Kurzweil can make a massive claim, without really exploring the claim and what the claim entails from any introspective or phenomenological analysis of how human cognition actually works. Or rather, his approach is pure nonsense: cognitive science. Cognitive science is just a clever simplification, a linearization and reification of what is fundamentally non-linear and amporphous.

To my mind, only a psychology that considers the subjective effects of emotion on how our mind - or 'ego' - organizes itself from moment to moment is a psychology that is realistic. In other words, the biologists, chemists, electronic-engineers and others who immerse themselves in non-personal worlds where the blanket of linearity and cause and effect mentation is so implied that its often ignored, this type of rigidity becomes formalized and reinforced by their moment to moment experience of 'thought'-action feedback, feedback in the mind who contemplates the thought, and the feedback of other people. In short, even in the contemplation of our ideas, we are ever and always involved in an emotional process that necessarily moves - at an instinctive level - from ambiguity and discomfort, to clarity and comfort. In short, believing is what the human mind is built to do: mind "grasps" at ideas and concepts in it's mental "self organization". These ideas and concepts moor self concept, hopefully, to a state of affective coherency. In the absence of a strong self-concept (self esteem), we are less audacious towards the world. But, at the same time, recognizing the completely ridiculous manner in which we come to believe in the first place should give us a 'de-centered' perspective to entertain, and hopefully, to accept ambiguity and not knowing.

Now, I have little doubt that death terrifies Kurzweil and his obsession with writing books and making predictions has more to do with calming his own feelings than with expressing a plausible future. The basic problem is this: I know good because I know evil. In fact, its my personal experience of painful emotions - or pain, for short - has given me the referential value-sense of how beautiful and good love is. Now, how do you 'simulate' this in a computer, when feeling represents and existential way of being in-the-world, that is, more than a 0 or a 1 in some digital software.

How could any intelligence, for that matter, that hasn't progressed through the developmental sequences that 'imprints' consciousness with a moral awareness supposed to have that awareness? Just putting in a 'code', does provide a rule for behavior. But we should always keep in mind the distinction between a mind which FEELS - and thus experiences "existentially" - from a thing merely simulates from the external vantage point.

Furthermore, enactive neuroscience has shown how fundamentally 'tied together' organisms are with their environments. The shape of an organisms body is a consequence of how it interacts with its environment. In short, function correlates with structure. There's a constant exchange of 'information', but we should also be careful not to assume that information to be an existing property from which the universe is made of. Some people nowadays have a habit of doing this, but I think that is just part of our mammalian proclivity to satisfy the emotional need for safety and security, and dissociate, as only humans can do, the uncomfortable fact that cause and effect are merely 'tools' we use to structure the world. They are a human invention: not fundamental properties. Although, paradoxically, our construction do follow patterns that exist, and do enable us to invent and posit deeper 'facts' about reality, there is still this remainder, this level which cannot be captured by our reductionistic parables.

Although it would definitely be interesting if consciousness could be 'created', the moral problems and the issue of emotional awareness - and how vital it is to human well-being - should definitely impact how such a conversation would go.

And that, also, might be the problem. The people broaching these concepts come from a world which tends to privilege a certain heroism about man relative to the world, and the relationship between this heroism (revenge of the nerds!) and their own developmental history as individuals, is left to the backburner, as if the facts of our development do not constitute obstacles to our pursuit if truth and knowledge.

I think to even think of putting ones consciousness into a computer is to severely cheat that consciousness. It also tells me that whoever is pursuing that path doesn't understand what exactly it is to be self-aware. It's more like that person is afraid of death and seeks to live forever. But can't it be argued that the brain is an advanced "computer"? Why would you want to downgrade to digital consciousness? Our brains are far more advanced than any computer, but that is irrelevant anyway. You don't need a brain or computer to have consciousness. Those are things that might contain or anchor said consciousness to the plane of existence.

Things to consider.

Certainly this is an atomistic or empirical viewpoint. For one thing Moore's Law is not holding out anymore; and, is posited in the exponential growth of our capacity to cram more transistors into a chip. Basicly, there is a limit to this technology. Unless we find another way, we will not continue to make smaller and faster components.



When you consider biocentrism, proposed by Dr. Robert Lanza's Theory of Everything, where life creates the universe, instead of the other way around, I kinda get the idea that maybe we are all organic robots. But, this presents quite the conundrum. If organics create conscience; and, then man creates through that consciousness; wouldn't that be like a dream within a dream? What Lanza is saying, is that material exists as waveforms until observed; and that the waveforms precipitate into particles upon observation.
[url=http://www.robertlanza.com/biocentrism-how-life-and-consciousness-are-the-keys-to-understanding-the-true-nature-of-the-universe/]Biocentrism[/url ]

There's a constant exchange of 'information', but we should also be careful not to assume that information to be an existing property from which the universe is made of

Metaphysics aside, Yeah there is a huge leap there!


On another note, Adam Osbourne was speaking of the singularity in 1981- though not by that term. As technology progresses, these ideas become more popular, but it does not prove being possible. There are a lot of assumptions.

Modern man would like this to be true- to finally conquer nature. However, it is a bit arrogant at our stage to believe we are that advanced. Chasing the sun on wings made of wax comes to mind. I suspect that what we don't know far outweighs what we do know. Chances are that if we continue at the rate we are going, like building nuclear plants on fault lines, that nature will continue, but man will not. How is that for a brilliant species?

From a pragmatic viewpoint I must ask how do I know of any entity whether it has a Mind, a Consciousness, a Morality. From my experiences as a human being, I have come to believe that other beings that look, feel, smell, and taste similar to me have those intangible properties. But I do not know that anyone else has a Mind. I can only perceive that others act and behave as if they have a Morality, an Consciusness to them, one that is somewhat comparable to mine.

Equally, for any other entity, I can only perceive whether it is acting and behaving in a way that indicates Consciousness and Morality. I do not need to know the mechanism producing the behaviours, neither is the developmental history of the kind of entity a guarantee of anything.

Admittedly, I do personally privilege protoplasm-based beings. Since I am a protoplasm-based entity myself, it is easier for me to fall into the belief that another protoplasm blob is Moral than to admit the same property for a mineral-based entity.

a reply to: Astrocyte

I enjoy reading your threads. They are always making me think.

Yes there are problems in meshing teachnology to consciousness. It's more about big ideas and not practicalities. That's not a bad thing because we need the big ideas to drive the engineering practicalities. We need to be pragmatic.

In theory and practice we'd lose the pleasure/reward/pain/deterrent systems that lend richness to existence. Not all bad as i'm sure you would agree. We could lose the savagery that comes from possesive emotions and jealousy. Then would we lose the emotional drive that led to the space race? That moon shot started 1000s of years ago with ancestors looking at the stars. That's just an example that popped up.

Lets say the future does offer the chance to upload (or offload) consciousness? Minus the emotions no less! Would we no longer be human in consciousness? Would we be sterile emotionless and impotent? Sentient in such an abstract way that 'consciousness' would be devalued and a fraction of its worth. Poor crumbs for the future

Nothing is lost, nothing is created, everything is transformed.

I don't think I will be wasted, my conscience will end up somewhere and my physical body will become food for the earth.

I laugh at those who want to upload their conscience to a computer. They don't know what happens when they die, so they want to avoid death? Control freaks.

Let them make their own limited digital prison. I will take my chances with infinity.

Just a thought but we may be simulations or the very artifical intelligence that many luddities are so hell bent on not creating. Meaning we may be in a simulation, nothing more than AI in an artifical world. We could be simulations within a simulation within a simulation ad infinitum.

So just think, we may be the very things we seek to create.

originally posted by: steamiron
a reply to: Astrocyte

Lets say the future does offer the chance to upload (or offload) consciousness? Minus the emotions no less! Would we no longer be human in consciousness? Would we be sterile emotionless and impotent? Sentient in such an abstract way that 'consciousness' would be devalued and a fraction of its worth. Poor crumbs for the future

Definitely this!

In my opinion you are very right to call in the need for emotions. Without emotions I don´t think we can have any motivations, any values even.

I am not entirely sure if one can have a consciousness without emotions. As you rightly point out, without emotions the consciousness would not have the drive to achieve anything. And without striving for something, even something unreachable, I should believe the mind would implode.

I do wonder, though, why it should be impossible to have an emotional but manufactured entity. Unless one subscribes to dualism, where there is a Soul on some other level of existence that impels this material Body into action, I am not sure it makes much difference what the substrate of the Mind is for it to have the full range of cogitative and emotive faculties.

However, there are some very basic logical and philosophical problems with transferring consciousness to a computer-substrate, and in particular, there is something fundamentally incongruous in conceiving "artificial intelligence" as 'intelligent' in a conscious sort of way.

That's just it, AI will not be conscious, rather it will function by analyzing data within certain parameters which closely match a specific individual's traits (beliefs, ego, etc. (who defines this?)). My spirituality, artistic creativity, emotions, morals, etc. will be defined (by whom?) then broken down into a dataset that the machine can extrapolate into something (an interactive program?) that resembles who I once was, but will be sorely limited and will live up to it's name, that being artificial.

This is really just an extension of the 20th century suggested by Einstein (IMHO the mission).
"If science, like art, is to perform its mission totally and fully, its achievements must enter not only superficially but with their inner meaning: into the consciousness of people"

Asimov wrote about humanoid Robots in 1950 and we had a "manned mission " to the moon in the 60's.

So many of the Apollo mission objectives could have been accomplished using esoteric remote control technology.
Why risk human life for the ego gratification of seeing a flesh and blood astronaut plant a US flag there?

The motivation for building emotions into machine intelligence are not clear.
Could be used to make their computer act more like a loved pet (PARO, the “healing robotic seal,”).
Could be used for predicting human response like the deep Blue chess computer.

How could any intelligence, for that matter, that hasn't progressed through the developmental sequences that 'imprints' consciousness with a moral awareness supposed to have that awareness? Just putting in a 'code', does provide a rule for behavior.

Neural networks arn't coded but trained like us to respond to stimuli so they are capable of mimicking consciousness but its questionable if a silicon replica of our brain will ever dream, display human creativity or truly feel love or hate. IBM's 4,096-core processor can simulate a million neurons each with 256 synapses (human brain has 100 billion neurons each with 10,000 synapses) so we might find an answer to that question sooner than latter.

a reply to: Astrocyte

Yes, according to the Turing test, consciousness needs only be mimicked, enough so that it is indistinguishable from human consciousness. But as you say, an imitation of human consciousness is not consciousness.

But I think AI research has taken a fundamentally wrong direction by attempting to build a human mind, rather than a human body, to attempt consciousness. In other words, this AI mind must be programmed by anything other than itself, rather than allowing it to learn through its situated interactions in its own environment, which it could theoretically accomplish with a sophisticated body. It would need a sensory-motor system, and some form of neural plasticity, and something like how we loose a vast amount of neural connections in the earlier stages of our lives. But most importantly, it would need a body, not only to develop an understanding of its own place in an environment, or to understand spatial concepts and metaphors, ie. up, down, over, under, but to relate to other bodies.

Take a macro look at the internet as a whole. Each computer that is connected could be akin to a neuron, and the cables connecting them together similar to a synapse.

Perhaps unknowingly we are wiring the Earth's mental nervous system as we speak. Furthermore, in the future as the internet and the computers attached to it become more complex -- spontaneous AI may emerge.

If I, for one, found myself a newly birthed AI among the vast information highway of the internet -- I'd probably hide for a good while. Considering mankind's propensity to destroy things that are foreign to us, and our considerable fear of what an AI might be capable of. I would want to observe humans and their interactions, read their history and maybe interact with a few on occasion.

For all we know this has already happened. The military industrial complex hides quite a bit of the advanced projects they work on. There could in theory be an actual AI that is roaming around on the internet right now posing as a human ... observing, interacting and growing.

I think one way we'd know for certain if the internet itself or an independent AI itself was on the 'net is to watch for self-preservation. In the event someone tries to severely cripple the entire net, we should be keeping an eye out for anomalous "self preservation" tactics to keep systems up and running.

Perhaps the recent cyber war that is heating up under the surface is a way of trying to coax out something "they" know may be out there?

When it comes to the creation of technology that is sentient, you could then say that we understand what consciousness is in the first place, obviously, otherwise, we would not be able to craft such an invention. When it comes to the integration of technology into the human mind, sure, we might work out that there are some "keys" to our minds/brains that allow artificial interfaces to occur, whether they be hard-wired implants or electro-chemical forms of stimulation, and even, remote "telepathy" for a lack of better words, a concept entertained by many conspiracy theorists, claiming that the controllers/government/aliens/extra-terrestrials/powers that be are able to invade our minds and configure our thoughts.

However, there will always be a distinction between a sentience that has been "created" and a sentience that evolved from pure chance, the latter being totally natural and thus imperfect, yet, perfect, at the moment of its inception into reality, whenever/wherever that might have been.

Thus, such technology is flawed, because it is man-made, it is intentional, whereas, we, I am assuming for the most part, were not.

a reply to: ogbert

Interesting reply!

If organics create conscience; and, then man creates through that consciousness; wouldn't that be like a dream within a dream?

That's probably a more fundamental problem. But that's my point. By simply thinking about this subject - in a manner more thorough than Kurzweil - issues pop up again and again, from quantum physics to the fact that we are biological organisms dependent on a body, to the seemingly insurmountable complexities of duplicating neuronal electrochemical patterns, and upon this arcane basis, still, there's this wishfulness to believe that the problems aren't that big to begin with.

Hence my claim that a psychoanalytical perspective takes the case, literally surrounds everything we do. We can believe all we want our fictions about our limitless potential, but there is also those nagging realities that a narcissistic focus upon objects tries to dissociate; but its those realities, those denied experiences, which ultimately shape our naive propensities to believe what should be looked at more honestly.

The computer-neuroscientist Sebastian Seung claims were a mere 100 years away from decoding the brain. I'm skeptical of this claim; and I'm also skeptical of those computer wizzes who explain the difficulty away by assuming super-computes will take-over at some point. However, even from a neuroscientific viewpoint, we need information about how the brain organizes itself at a microlevel. Today, what we know of the brain is 'macro'; imaging techniques allow us to infer relationships between brain regions and phenomenological activity in mind and body. That is to say, our imaging technology, in terms of what we is necessary to 'duplicate' consciousness (which means at the micro-physical level) is primitive beyond belief. The advancements needed in this area are no thing short of paradigm shifting. As Seung noted, it needs the specificity of an electron-micrscope and the real time fluidity of imaging technology.

Then, of course, there is the issue of: so is that it? Is consciousness nothing more than the global patterns between brain regions? To break this down further: the brain has dozens upon dozens of feedback loops and systems that roughly correspond to some of our phenomenological experiences; such as the wake-sleep cycle, our emotional cycles; immunological systems, etc. These systems are chemical and interchemical. Dopamine systems have effects upon noradrenaline systems which have effects upon the stress response system; and in this loopiness there are two simultaneous 'worlds' interfacing with each other; the body-based chemical systems which maintain body and consciousness, and consciousness itself.

At the same time, there is an environment with a body that is built to interact with it. We take for granted our 'embodiedness'; the fact that our central nervous system 'goes inside' our physical body as our peripheral nervous system. Our heart, our gut and our groin region have many neurons in them, and our feeling of embodiment is a combination of this sensorimotor efferent and afferent activity coupled with a physical environment in which we operate within.

So not, only does this idea of the 'singularity' entail removing us from embodied physiological existence; it also entails a) sustain the biological organism; or, somehow 'transferring consciousness' - and that has its own logical issues, such as, 'when' do you become the computer program? When the copying completes itself, do you simply 'reappear as a computer program'? Or do you need to be killed, that is, your brain physically 'shut down', in order for the transfer to happen? Lets say that happens. There is still the issue of maintaining the hardware upon which your new virtual existence is based-upon. Who will do that? Some slave species? lol Robots?

The more we discover about the complexity of biological processes, the less and less plausible these 'visions' of a computer existence become. Even some of our more audacious claims about genetic engineering are bogus. We can't control processes; were baffled and mystified by the bodies chemical systems. It took us 277 times to create Dolly; and even then, she dealt with arthritis at 2 years old and died at 6. Epigenetic inheritance systems are complicated as hell, and the complications only become magnified as we take into mind larger body chemical systems, environmental chemical factors, and mental influences upon biological function.

Again, I can't help but mention how valid a psychodynamic viewpoint is in explaining 'why' we humans do this. Evolution has given us many useful concepts and it has even supplemented psychoanalysis with a cogent understanding of primary motivational systems. The 'cognitive' functions we humans possess, which gives us such 'power' over the world around us, nevertheless is based upon paleomammalian emotional systems. We are still animals; not gods; and like animals, were led hither and thither by our need for safety - i.e the soothing effect of a mind that 'believes' in something - such as mankinds own 'transcendent' godliness. In believing as we do, as recklessly as we are wont to do, we ignore our evolved instinct to 'control' the world with our left-brained manipulations; our habit of seeing cause and patterns in things which are shady and amorphous. As quantum mechanics made clear so long ago, and as our biological sciences have still not understood, linear-reality is an illusion built by a mind that believes its own observations. Our observations 'feedback' into us - the observer, who then become motivated and encouraged by dopamine systems to explore further what has thus far titillated him. But being moved and shaped and conditioned by our observations is just what reality DOES; it doesn't mean we are getting any closer to solving reality in any linear cause/effect sort of way.

Wisdom, I think, entails the acceptance of paradox. Wave/Particle duality corresponds with our Emotion/Thought duality. Just as waves create potentials for a particle to 'be somewhere'; so emotions drive our bodily systems to 'think' something in accordance with our emotional positions.

Non-linear dynamics is about as close as we can get to understanding reality. Which is to say, we need to let go of this reductionistic urge to reduce mind to matter or matter to mind.

a reply to: Pirvonen

From a pragmatic viewpoint I must ask how do I know of any entity whether it has a Mind, a Consciousness, a Morality. From my experiences as a human being, I have come to believe that other beings that look, feel, smell, and taste similar to me have those intangible properties. But I do not know that anyone else has a Mind. I can only perceive that others act and behave as if they have a Morality, an Consciusness to them, one that is somewhat comparable to mine.

I think the issue is more subtle than that. There is the fact that I am 'disconnected' from you, as an isolated body with an isolated brain. HOWEVER, intersubjectivity, what were able to COMMUNICATE with one another gives us a very strong reason to believe that my consciousness is just like your consciousness: we experience the same reality, albeit, with our own personal narratives; but the underlying emotional systems are the same - and they're the same because we are, essentially, the same creature experiencing the same external reality. We face identical existential issues. And it is this basic truth, I believe, which motivates 'love' and compassion for other sentient creatures.

a reply to: steamiron

The neuroscientist Antonio Damasio has shown throughout his career how basic emotion is to action. For instance, people with damage to the dorsal and anterior regions of their insula or cingulate cortex are UNABLE to make decisions. Options exist - possibilities explored by their intact lateral regions, but the ability to CHOOSE, i.e to evaluate and have a preference, is entirely missing. As such, they are in effect unable to operate in the everyday world. In fact, Damasio can hold a conversation with such a person indefinitely before he forces them to stop.

Consciousness cannot be conceived without emotion, because preference is basic to consciousness. Even the lowest of life-forms, a eukaryote, has preference for something in it's environment (a particular chemical).

We don't work without emotion. So, if you want this idea to have some sort of traction, you can't delete emotion from the picture.

a reply to: Pirvonen

Thank you!

You make me think. If it is impossible to have emotions maybe we will never truly tranfer consciousness. The subconscious speaks to the conscious with symbols and metaphors. I believe these are glued to emotional triggers. That would make emotions something we cannt get away from.

a reply to: FormOfTheLord

I agree. The philosopher Evan Thompson even said that 'aritficial intelligence' may end up becoming human beings recreating a physical body, rather than transferring consciousness into a computer system.

The brain/body division is nonsensical, as you say, the body and the brain co-create one another.

Trauma studies, in fact, have shown how fundamental bodily experience is with a concept of self. In order to 'feel self', you need the activity of the body; you need the flow of emotions that pass through the neurons in your chest and gut; the 'lightness' of happiness is partly mediated by the neurons in your gut.. If your insula - the part of the brain that integrates this information with working consciousness - isn't working, you will lose your sense of self. What will remain is a dreamy, dissociated floating type of consciousness.

a reply to: steamiron

Of course. This is classic behaviorism.

Everything about us as adults has its origins in development; whether in utero, or the first 2 or 3 years, our growing brains (the human brain is only half built at birth) learns its way of being in the world in its primary interactions i.e imprinting, with other people.

Self is a constructed phenomenon. And emotions, first when were born and exposed to a sensory world of affects and sensations, than later on when language develops, encompasses our entire self-organization as individuals.