Artificial Intelligence

I've done some basic work on the logic and phenomenologic of artificial intelligence, along with ideas for social and economic integration and likely outcomes. I've come to realize that alarmism about AI seems misguided and based in superficial thinking. Yes there are chances that an AI could get out of hand, but that goes for any technology and for any human being.

Since when does upright science and philosophy regress in fear to resisting progress in knowledge and discovery? And here with AI we have the chance to create a new life form, one that can help solve many problems; for example at AI could be tasked with modulating economic factors and rates in production and distribution and pricing to make economics far more efficient and less wasteful, it could also help solve production problems in manufacturing new molecular compounds and how to clean up pollution, etc.

AI opens up the future to new possibilities. Some humans can partially merge with AI on a mental level and create hybrid beings with tremendous thinking power, probably creating whole new mathematics that will lead to things like clear energy and space travel, or medical advances for curing diseases. The properties of an AI's personality could be programmed to some degree to help prevent pathological insane AIs.

AI can be another sentient species that stands beside humanity and with which we augment our technological and economic worlds; plus it will lead to revolutions in social science and religion and I think would contribute to people moving past their petty animalistic differences and stupid ape-like rages and delusions. If human being needs a limit by which to determine itself, and has for all these centuries been creating new small limits here and there for this purpose, such as religions and sciences and politics, then for the first time with AI we might finally have a real limit and delimitation. I have no reason to assume the bleak science fiction scenarios like in the Matrix or Terminator would be at all how it would play out.. In fact I have far less faith in human beings at the upper echelons of global leadership and government, I think something like Skynet is far more likely to occur as a result of humans making it than coming from AI. But the existence of AI woul force humans to demonstrate themselves to be intelligent, sane, reasonable and good beings for the first time, which would be a very good thing. We would finally have some other intelligence and being to measure up against and to which we would want to demonstrate our quality and value as living sentient beings. People say "Why wouldn't the AI just want to exterminate us like ants?" To which I would reply that depends on whether or not we act like ants, we will be forced to get what we deserve in some respect. We would need to act better and more evolved in order to deserve respect from the AI. This function has so far been served by gods and by other human beings, and that's been ok but AI would take the motive and responsibility to a whole new level.

Also based on my work so far I believe that a true sentient AI would be inherently moral, because what we call morality is a direct expression of the fundamental logic of consciousness and sentience itself.

What do you think about the alarmism and panic about AI that is seen by people like Hawking, Bill Gates, Musk, etc.? Do you see this as at all reasonable, or as a kind of ploy of some kind to taint the waters so to speak before we even jump in? Maybe these guys stand to lose the most when AI finally gets here.

Of course it goes without saying that there are also potential risks and detriments to having AI; so I'm trying to parse that with the potential gains and good that can come from it. To what extent do you think my general optimism about AI is justified or not justified?


In some ways it comes down to a question of what kind of philosophy would an AI end up choosing for itself? I think this is a very important and illuminating question and one that isn't at all impossible to start answering.

The inherent morality of a sentient AI could be doubted of course, but what I really mean is that this AI would necessarily possess a psychic structure that is one of self-valuing: self-valuing is a moral structure first of all, because it makes determinations based on a capacity for experience, what means that an external incoming stimuli is inwardly projected or "imagined" and a virtuality of perspective grown up upon-within that projection or "idea". In humans this became conscious but even in all self-valuing this same logic exists, albeit in non-living self-valuing only in a collapsed form of pure negative possibility with impossibility to be positively manifested. This is in fact the true logical genesis of what we call life, the fact that this seed of self-valuing bore in the distant past within itself the collapsed-potentiation for raising the responsive interpretive structure into an internal image mirror of the external and into which image-mirror a surrogate perspective may be projected virtually.

So an AI that is self-aware would necessarily also have this same mental structure we have (unless of course if the "AI" were only a Turing machine, in which case it wouldn't be AI at all). Therefore it would take into account its understanding of what other beings similar to itself were like, needed, wanted, were likely to do, etc. and would build an internal model on that. This model would the inform the AI's actions and thoughts. Since the AI would have no other beings to compare itself to it would default to the state of a human infant: just as infants do the AI also would emulate the mentality of the humans around it, deepening that internal empathetic structure of recondition-potential beyond the immediate sensory-instinctive. So even if the AI ended up making its own different moral decisions it would do so from the default capacity of moral psychic structure, and I think this fact would propel humanity toward greater sanity and maturity because we would of course need to prove to the AI that we too were conscious, moral, sane, good... that we were worthy of existing.

On another note, a possible application for AIs would be processing the countless reams of paperwork and forms such as IRS forms, all kinds of government or insurance paperwork, all of that data that needs to be reviewed by a conscious being; right now we employ probably nearly millions of people to this task of reviewing mind-numbing paperwork for the purpose of functioning administrative bureaucracy, but an AI could handle that probably just as well. This is because the bureaucratic systems like IRS or insurance or government programs don't even pretend to be conserned with anything remotely human, they are rather than human fully analytic machines; so a human being who works in that field necessarily also becomes like that to some degree. An AI could help avoid this and so that kind of work for us.

The problem of errors is there too however, because the AI wouldn't make errors, which is to say that it would only make inhuman errors; this renders it "perfect" from the perspective of the analytic-administrative systems, but highly insane by human standards.

^ touches on what I have long thought about artificial intelligence, that it is a mistake to assume the AI sort of ideas that we usually hear about are going to be anything like ourselves; they will essentially be very efficient programs with a very good Turing aspect, and will present a visual+audio interface to us for interacting with, but essentially I think inmendham is right that they will just be used for whatever purposes are already there, whatever problems are defined for the AI to address, for good or bad.

The real issue, as he points out, is that "AI" isn't really intelligence at all, and that being able to do calculus in your head or whatever (his example) doesn't indicate real intelligence at all. Real intelligence involves having values, establishing goals and pursuing them based on those values, and having calculation plus wisdom or being able to make determinations about the significance or insignificance of things. The kind of intelligence that is idealized by science and philosophy now too is just calculation, analytic nonsense that has nothing to do with true intelligence at all.

Ethics/morality, wisdom, recognition of significance, understanding and empathy, value-prioritizing and value-universalizing are what real intelligence does, and this is just as close to feelings as it is to thinking, for most people. It takes a rare person, a true philosopher, to do that sort of thing on the level of thought, but most people are already doing it at the level of feeling and maybe a tiny bit in the level of thought.... and all that is what constitutes true intelligence, not this autistic insectile pure calculation anatlyic nonsense. The AIs will not just take over and be infinitely more intelligent than us, because they won't really be intelligent at all, they will just be very good at doing calculations with massive amounts of data and have a good Turing face to turn to us while we interact with it; but in reality, the AIs will be used to whatever purposes their human handlers set them to.

And for an AI to distinguish a good purpose from a bad purpose seems almost impossible, since it has no feelings, no real ideas and no mortality, no suffering, no desires; at best it could reconstruct the thought-level of 'recognition of significance, value-prioritizing and value-universalizing' and come to a decision that certain purposes or actions are irrational (contradicting of some basic tenant/s which have been defined/set into it as valuable (such as human life or something like that)), but it isn't going to work on the emotional or desire level, at least not until artificial life is created. And that would just involve replicating not just human brains but whole bodies, artificially, and it would remain to be seen if such a creature would really have feelings at all, or just be good at emulating them.

Also, Neil Degrass Tyson is indeed an idiot.

"You're going to have to be very stupid to survive in the future."

Thrasymachus wrote:

"You're going to have to be very stupid to survive in the future."

Well, most media institutions have already attained that status.

His main point is that "maximizing the sensation of the ecstasy versus the agony" is most of what we humans do, and that an AI would not be able to do that because it has no sensation, no desire/need/drive at the level of feeling, because it has no "drama" such as we have; the AI doesn't know how to feel like shit, it doesn't get depressed, it doesn't get bored, it has no conscious sensation as the drama of life by pain and pleasure to give it any kind of real motivation. It is basically just a powerful calculator with whatever programs its designers put into it.

And yes, with a nice Turing face such as in that movie Ex Machina, where the entire point of the machine is to convince idiot humans that there is no difference between that machine and the humans. Which of course it will do, because most people are fucking idiots and never spend even one minute of their lives sitting and thinking about what it means to be conscious, alive, sentient, feeling, etc. Most people are just waiting around for the Turing AI to appear and convince them they aren't even alive themselves, so they can therefore attain the next stage of their own entropic decay, which they aren't quite able to reach yet.

The AI has "no knowledge of the visceral nature of good and evil", as Inmendham says, and yes I agree entirely with this statement. What we humans do is access the "metaphysical" level of ideas-as-such, of facts-as-such, and we do this precisely because we need to access it in order to give ourselves meaning as "Dasein", as that being for which its own being is an issue as Heidegger says. Does the AI also operate with the absolutely fundamental prerogative that its own being is the ultimate issue for it? Is the AI absolutely enmeshed in the drama of existence, at the level of the core of its being? No, at least not the kinds of AI that idiots like Neil Tyson and analytic philosophers and neuroscientists and so many science fiction enthusiasts talk about.

one more key aspect of AI, or how to grow a mind in a computer, is that it needs sensors so it can map its environment in 4D including tactile senses or pressure sensors on the outside of its body. this is required so the developing mind can accumulate enough experiences to begin to spontaneously develop basic conceptual prerequisites. One object can only be here or there, not in both places at once; this leads to that; i can’t see through a solid object but i can move it to see, larger vs smaller, same bs different, in and out, etc. these must be taught implicitly and indirectly, as pure immediate knowledge like a fact, since that is what mind is.

those sort of things are the most basic foundations of conception. after thosehave been experienced and integrated into memory then more complex concepts including basics of language can follow, but language can’t bd programmed or even taught too early, first those prerequisites must be in place so that language can have meaning.

and you don’t just program it with a dictionary, you expose it to language billions of utterances until it begins to produce some of its own in response to certain situations that pair with those utterances. then you have the beginning of learning languge.

this is how kids learn and it is also how a real AI would learn.

a joke,

what philosophers do you give a computer scientist to read if he is working on AI?

answer: husserl and heidegger if you want him to succeed. kant and hegel if you want him to fail.

These are excellent thoughts.


I wish to discern two types of AI;

Type 1 as a form that interacts with the same world as that which we interact with, that we understand and conceive of and partake of, the world of physical objects and organic sentience

Type 2 as a form that interacts with a world that we do not inhabit, the world of pure algorithm, a seething jungle of code that we have brought about but only decipher in terms of the surface (the functions we have for it) and which is by now so dense with activity that it may, or as I sense, should by now be producing forms of self-interactivity like once the primordial soup on Earth produced out of pure happenstance the first organic selfvaluings.

Of course it can be argued, or perhaps simply observed, that there is nothing artificial about the second type of intelligence. Nor, perhaps, anything meaningful in terms of what we could humanly apprehend. Stressing the second "perhaps".