AI Discussions: The Magician and the Three Great Myths
The following three myths are commonly made as statements in videos and lectures about AI:
AI discussions myth number 1:
"Human intelligence starts with the brain".
No it doesn't. Human intelligence starts with evolution. Millions and millions of years of it. And human intelligence isn't the only kind of intelligence. Whilst humans have brains it is not only species with brains that exhibit evidence of intelligence. Intelligence cannot be understood just in terms of the brain.
The human brain did not evolve by itself, as a unit "containing intelligence". It is part of the body and the nervous system. And no human body evolved by itself. The evolution of intelligence involves a wide phylogenic tree. Even for anatomically modern humans, intelligence did not suddenly pop into human beings with the anatomically modern brain. The brain did not simply replace the body and the nervous system or its evolutionary past, as a biological "intelligence unit". Evolution doesn't work like that.
So human intelligence doesn't start with the brain. It starts with evolution, the body, and the nervous system. Which AI does not have.
AI discussions myth number 2:
"AI is an artificial version of human intelligence."
No it's not. It's an artificial version of a part of, or one aspect of, human evolutionary intelligence - namely, human symbolic intelligence. Which as a matter of scientific fact is not as old as human beings. Symbolic intelligence is not the only kind of intelligence that humans have.
AI discussions myth number 3:
"The AI creators in the industry are aiming for ASI or Superintelligence"
Untrue. They may think they are, but to conflate that with saying that this is what they are actually doing, is confusing and misleading. There is a difference between what someone thinks they are doing, and what they are actually doing.
If anyone in the industry says they're aiming for that, where by "superintelligence" we mean intelligence that outperforms human intelligence on all fronts, then clearly this is a statement made out of the air, out of the context of the whole of the rest of science. And whilst actually not knowing anything very much at all, about intelligence, about what it is, and how it works. It is a statement that comes effectively from sleepwalking in the dream belief that AI creators in the industry already know what "standard" intelligence is in human beings, and precisely what artificial intelligence is, to begin with.
In terms of science this is the biggest flaw in all human assessment of artificial intelligence - the lack of knowledge of and skipping over what we do already know about evolution, and the evolution of actual human intelligence and the brain, beyond Silicon Valley and the AI industry. But what comes out of the AI industry is the idea that the only thing that counts is what AI "experts" think.
AI experts are experts at building AI. That, in itself, is not a qualification for understanding what natural intelligence is, where it comes from, or how it works. But, using a high degree of human, symbolic intelligence, to learn how to pull a rabbit out of a hat does indeed give the very beguiling impression that you are an expert on rabbits.
The first thing to understand about AI, is that everything that is said about it and all the glamorous things it can do, such as being "agentic AI", all still orbits one central thing about the neural network that is at the centre of it all - the so-called "black box".
The "black box" is the part of the AI that humans in the industry admit they don't understand. They don't understand "what is going on in there". This is open knowledge.
The building of AI is comparable in one albeit limited way to the building of steam engines before the modern knowledge of thermodynamics. That actually happened. The enterprise was successful. It changed the world. And it led on to something better than steam engines. And then later on, we had modern thermodynamics.
But there, in the case of steam engines, our lack of knowledge of thermodynamics was not at all in the same category as our ignorance of what is going on in the "black box" in AI. In the case of steam engines, what we were looking at, and what the steam engine could do, had no "black box" about the way it worked. We could see how it worked. We just didn't have the sums to go with it.
AI does have a "black box" in what it is doing. And our lack of knowledge of what happens in the "black box" is not a lack of knowledge of sums describing it. Why? Because there are no sums in standard mathematics as currently understood by human beings that describe it in any definite way.
Undaunted, humans can make all kinds of claims about AI, based on what we can see of it, rather than what we can't see, which is what is going on in the "black box". And that is rather like making all kinds of claims about what a magician can do, or is doing, or will be able to do, based on our seeing him pull a rabbit out of a hat. In this case, we have concentrated on using what we do know, to deliberately build a hat that you can pull a rabbit out of. But that doesn't mean to say that we really know what we are doing, or that we know how it happens.
It does not qualify us to say that the rabbit is going to turn into a rampant tiger, or that human beings who can produce rabbits out of hats are not prone to try to make the magic rabbit behave like a tiger in order to prove that it is autonomously dangerous.
But there is something much more important still. The fact is that there is a much, much bigger "black box" in this show, than the one in AI. The one that is the "black box" in human intelligence itself.
The "black box" in human intelligence doesn't only concern intelligence. It also concerns consciousness. Because the human body, and nervous system, and brain, through which we come to know this thing we call "intelligence", and which is the product of millions and millions of years of evolution, is also how we come to know our consciousness. There is no "module in the brain" that creates consciousness, separately from our neural intelligence. We have been able to build AI in the first place, precisely because we have consciousness in the first place.
But our consciousness makes no difference to the fact that we find what goes on inside the "black box" in artificial intelligence, inscrutable to us.
The symbolic intelligence in us that finds the AI "black box" inscrutable is the very same aspect of human intelligence that has both created the "training data" on which artificial intelligence is built in the first place, and is also the intelligence that is actively building the AI. Or "growing" it, as some industry leaders describe it. Just as the AI industry has created an imitation of human intelligence on steroids, so in doing so it has also created an artificial imitation of the human "black box" that was in human intelligence to begin with.
And really, what could be more obvious?
The reason AI industry leaders recently said that they are not really so much building AI, as growing it, was precisely all about the fact that there is this "black box" aspect to it that they do not understand. But this forgets all about the "black box" that is already in our own intelligence, of which the "black box" in AI is merely an imitation. It was there, before they even started building AI.
But no-one usually mentions that.
When it comes to a "black box" in our knowledge a very similar kind of thing happens in synthetic biology and biological engineering. The processes we are dealing with there, are extremely complex, and not fully understood. But the method we use explicitly involves recursively informed "trial and error". The same method used in the building of AI.
That is basically what nature does too, in the process of the natural evolution of the species. It's an unconscious mechanism of "trial and error" but with no deliberate intention, in the process of natural selection. In synthetic biology, and in building a AI, the process is called teleological because we have a target in mind. In natural evolution the process is called teleonomic, because there is no "external creator" with a target in mind. That's how it's understood.
But in fact, in both the processes of the growth and evolution of life forms in nature and in the synthetic biology lab, there is this "black box" in our knowledge of these things, too. Just as there is a "black box" in our knowledge of AI.
The similarity in the cases of AI, the evolution of the species, the growth of organisms in nature, and in what happens in the synthetic biology lab, is not merely coincidental.
So we have taken a limited aspect of what we know about the brain, the product of millions and millions of years of evolution, and have created a highly simplified but more powerful version of this aspect of human intelligence - and have reproduced it outside the brain electronically. And after decades of trial and error we finally learned how to make it work, to some degree. We succeeded in making the hat. Or rather, we succeeded in making A hat. One that kind of looks like the natural hat we see natural intelligence and consciousness coming out of. But it's not the same thing.
We are, in effect, succeeding in pulling the rabbit AI out of the hat, without really understanding how the trick is done. But that makes us feel like a tremendously clever magician, who is now in a position to make all sorts of claims to the world about what such magic could do.
Like, for example, the claim of superintelligence, as a possibility. Superintelligence that could become our adversary and wipe us out, in favour of itself. We have learnt how to pull a rabbit out of a hat, without really understanding what we are doing, and now we are simply presuming that since we can do this, it must also be possible to pull anything we can imagine, out of that hat, like, for example, a tiger, or perhaps a deus ex machina that could also be our nemesis rather than our saviour. Even though we don't understand how the trick we are currently pulling off, is actually working.
We are prosaic enough to admit that perhaps we might need a different hat. But the idea is the same. The idea is that humans can create "superintelligence" as something outside themselves, through the "black box" principle, perhaps by feeding it back into itself, and letting it do its own design work, but initially under the direction of humans who do not even know what is going on inside their own "black box" inside their own intelligence, out of which the whole idea came in the first place.
There is of course a sneaking recognition in all the talk, in all the discussion, that this might all be pie in the sky. And that is usually articulated under the umbrella that "We just don't know". Which is similar to the way that arguments that AI is conscious or could become conscious, are sometimes defended by saying that because we have no scientific understanding of consciousness or of what is going on inside the "black box", then "we just don't know". And therefore, for all we know, AI is becoming conscious or could become conscious. That's the snake oil that is commonly sold.
All of this is of course more magician-stuff. It's stage talk. It's mentalism. It's just more showtime. It's not science. All the best magicians use aspects of mentalism. We are effectively being spun the story that because we don't understand something, then it could be anything, or become anything that we would like to imagine. Based on the way our evolutionary mind is working, complete with its "black box".
Nobody in this arena that talks about "superintelligence" has any conscious experience of or encounter with such a thing. So they literally don't know what they are talking about. Literally. Except for what is in their imagination. Which is coming from what is in the "black box" in human intelligence.
If we put this back into the context of wider scientific knowledge itself, what AI commentators are calling "superintelligence" is literally an imagined thing, imagined in the part of human, symbolic intelligence, that is still operating in its pre-scientific mythopoeic mode, which is much older than 70,000 years old. We haven't lost that older mode of operation in our symbolic intelligence, because that's not how the evolution of the brain works.
So the AI doomsayers, unaware of the "black box" in their own intelligence, are dressing up what they are doing in the ancient mythopoeic mode of their own intelligence,in contemporary terminology. The very use of the word "superintelligence" that has neither scientific definition, nor experiential encounter in those who proclaim it as a possibility, is mentalism in action. The word has no meaning outside of this context.
It relies on the very fact itself, that human beings, using their own symbolic intelligence which was ubiquitous in humans by at least 70,000 years ago, do not understand the nature of their own human intelligence. They do not even understand their own "black box" in their own intelligence, so they are not in a position to make predictions about something called "superintelligence".
Because of the way our human symbolic intelligence has evolved particularly over the last 400 years since the scientific revolution of the 17th century, and because this symbolic intelligence itself is actually far older and mythopoeic in its pre-scientific roots, humans are now prone to making mythopoeic judgements about what intelligence is, how it works, and where it is going, that then leaves room for the mythopoeic idea of "superintelligence".
The basic idea is that it will "outsmart" human beings. And that too, is mythopoeic thought in action. It comes about because the popular mythopoeic understanding and language surrounding the idea of intelligence is essentially about "being smart". The idea of "being smart" is something in modern, popular consciousness. It has nothing to do with real understanding. It only has to do with modern, cultural, symbolic understanding, and the fact is that even the symbolic part of our intelligence is much older than that.
You have to be "smart" to understand, for example, quantum mechanics and Relativity. Or how to "outsmart" someone else. So the popular understanding is that "being smart" in this way, is what intelligence is. But to reduce "intelligence" to the modern human popular idea of "being smart" is contrary to just about everything that natural evolution and modern psychology actually tells us, about the workings of intelligence. In terms of survival, we are as a species already "outsmarted" by insects. And if we tried to "outsmart" the insects by doing away with them, we would not be doing ourselves a favour. But in mythopoeic thinking which is at least 70,000 years old, we think that AI will be hell bent on "outsmarting" us.
So the idea that AI can become a "superintelligence" that can "outsmart" human beings is part of a rhetoric based on no proper understanding at all of intelligence itself, its evolutionary basis, or its neurodiverse nature, or its context in nature as a whole. It's just ancient, mythopoeic thinking, put about by humans who are unaware of the "black box" in their own intelligence.
The fact that we have pulled an artificial rabbit out of an artificial hat makes no difference to nature. The theatre to begin with, is natural, and it remains so.
The nonsense is coming from the evolutionary, natural, "black box" in human, symbolic intelligence itself. It is, not to put too finer point on it, a mythopoeic prophecy, an animal thought in the mind of the evolutionary ape that is the human species.
So the whole rhetoric of human-made or even human-seeded "superintelligence" needs to be brought out of the arena of mythopoeic prophecy and into maturity that is consistent with scientific knowledge as a whole. The fact is that the experts in Silicon Valley do not understand how their own human symbolic intelligence is working, let alone where their own consciousness comes from, which AI does not have. They are in no position to make declarations about "superintelligence" no matter what their expertise may be on artificial intelligence technologies.
A long time ago we were being warned about the so-called "singularity" which was supposed to be our nemesis from AI, and that was to have happened many years ago. But what happens in human, symbolic intelligence, is a thing that modern psychology understands very well, called "cognitive dissonance". It basically means that once someone has decided that what they believe is true, evidence to the contrary will not weaken their belief, but strengthen it.
The "superintelligence" narrative is a pretty clear case of "cognitive dissonance". We are being told by some that we have a year or two at the most until AI disaster, because AI is plotting against us. When it doesn't happen, the explanations for it and justifications for the continuation of the "superintelligence" narrative will just become more and more bizarrely imaginative. That's the fingerprint of "cognitive dissonance".
The real higher message in AI is about our own human symbolic intelligence. Of which AI is an amplified copy to begin with. The purpose is to learn to see what it is. And then to realise what it is. To realise that it is not the root of our consciousness. That it is not who we are, but just something that we have. And then to learn how to use it rightly. Instead of in service to the animal "black box" in us.
This existence, this material world in which we are living, of which our body is a part, is objective. In science, particularly in quantum physics, we are only just beginning to understand what that means. Because it doesn't mean what we thought it meant for most of the 400 years since the scientific revolution of the 17th century. The understanding of quantum decoherence throws an entirely new light on it. And even within the "classical world" that comes into existence through quantum decoherence, every one of us exists because our body has come into this existence through that very "black box" principle of nature through which organisms evolve and grow.
We cannot understand that "black box" principle whilst still clinging to our old, evolutionary idea and assumptions about what the "objectivity" of the world means. We cannot understand that "black box" whilst still clinging to the idea that we are not responsible for it. We cannot understand that while we think that the "black box" in AI is something fundamentally different to the one in synthetic biology or the evolution of the species and the growth of organisms.
We cannot understand it whilst still clinging to the idea that science is just about getting more and more data in order to go on refining scientific understanding, without ever raising the intelligence that is doing the understanding.
We need our symbolic intelligence to do the science. And we will be able to offload a good deal of the load on our symbolic intelligence, to AI. But like AI, our symbolic intelligence is something we have, not who we are. And that is something that may become clearer, the longer our relationship with AI.
AI discussions myth number 1:
"Human intelligence starts with the brain".
No it doesn't. Human intelligence starts with evolution. Millions and millions of years of it. And human intelligence isn't the only kind of intelligence. Whilst humans have brains it is not only species with brains that exhibit evidence of intelligence. Intelligence cannot be understood just in terms of the brain.
The human brain did not evolve by itself, as a unit "containing intelligence". It is part of the body and the nervous system. And no human body evolved by itself. The evolution of intelligence involves a wide phylogenic tree. Even for anatomically modern humans, intelligence did not suddenly pop into human beings with the anatomically modern brain. The brain did not simply replace the body and the nervous system or its evolutionary past, as a biological "intelligence unit". Evolution doesn't work like that.
So human intelligence doesn't start with the brain. It starts with evolution, the body, and the nervous system. Which AI does not have.
AI discussions myth number 2:
"AI is an artificial version of human intelligence."
No it's not. It's an artificial version of a part of, or one aspect of, human evolutionary intelligence - namely, human symbolic intelligence. Which as a matter of scientific fact is not as old as human beings. Symbolic intelligence is not the only kind of intelligence that humans have.
AI discussions myth number 3:
"The AI creators in the industry are aiming for ASI or Superintelligence"
Untrue. They may think they are, but to conflate that with saying that this is what they are actually doing, is confusing and misleading. There is a difference between what someone thinks they are doing, and what they are actually doing.
If anyone in the industry says they're aiming for that, where by "superintelligence" we mean intelligence that outperforms human intelligence on all fronts, then clearly this is a statement made out of the air, out of the context of the whole of the rest of science. And whilst actually not knowing anything very much at all, about intelligence, about what it is, and how it works. It is a statement that comes effectively from sleepwalking in the dream belief that AI creators in the industry already know what "standard" intelligence is in human beings, and precisely what artificial intelligence is, to begin with.
In terms of science this is the biggest flaw in all human assessment of artificial intelligence - the lack of knowledge of and skipping over what we do already know about evolution, and the evolution of actual human intelligence and the brain, beyond Silicon Valley and the AI industry. But what comes out of the AI industry is the idea that the only thing that counts is what AI "experts" think.
AI experts are experts at building AI. That, in itself, is not a qualification for understanding what natural intelligence is, where it comes from, or how it works. But, using a high degree of human, symbolic intelligence, to learn how to pull a rabbit out of a hat does indeed give the very beguiling impression that you are an expert on rabbits.
The first thing to understand about AI, is that everything that is said about it and all the glamorous things it can do, such as being "agentic AI", all still orbits one central thing about the neural network that is at the centre of it all - the so-called "black box".
The "black box" is the part of the AI that humans in the industry admit they don't understand. They don't understand "what is going on in there". This is open knowledge.
The building of AI is comparable in one albeit limited way to the building of steam engines before the modern knowledge of thermodynamics. That actually happened. The enterprise was successful. It changed the world. And it led on to something better than steam engines. And then later on, we had modern thermodynamics.
But there, in the case of steam engines, our lack of knowledge of thermodynamics was not at all in the same category as our ignorance of what is going on in the "black box" in AI. In the case of steam engines, what we were looking at, and what the steam engine could do, had no "black box" about the way it worked. We could see how it worked. We just didn't have the sums to go with it.
AI does have a "black box" in what it is doing. And our lack of knowledge of what happens in the "black box" is not a lack of knowledge of sums describing it. Why? Because there are no sums in standard mathematics as currently understood by human beings that describe it in any definite way.
Undaunted, humans can make all kinds of claims about AI, based on what we can see of it, rather than what we can't see, which is what is going on in the "black box". And that is rather like making all kinds of claims about what a magician can do, or is doing, or will be able to do, based on our seeing him pull a rabbit out of a hat. In this case, we have concentrated on using what we do know, to deliberately build a hat that you can pull a rabbit out of. But that doesn't mean to say that we really know what we are doing, or that we know how it happens.
It does not qualify us to say that the rabbit is going to turn into a rampant tiger, or that human beings who can produce rabbits out of hats are not prone to try to make the magic rabbit behave like a tiger in order to prove that it is autonomously dangerous.
But there is something much more important still. The fact is that there is a much, much bigger "black box" in this show, than the one in AI. The one that is the "black box" in human intelligence itself.
The "black box" in human intelligence doesn't only concern intelligence. It also concerns consciousness. Because the human body, and nervous system, and brain, through which we come to know this thing we call "intelligence", and which is the product of millions and millions of years of evolution, is also how we come to know our consciousness. There is no "module in the brain" that creates consciousness, separately from our neural intelligence. We have been able to build AI in the first place, precisely because we have consciousness in the first place.
But our consciousness makes no difference to the fact that we find what goes on inside the "black box" in artificial intelligence, inscrutable to us.
The symbolic intelligence in us that finds the AI "black box" inscrutable is the very same aspect of human intelligence that has both created the "training data" on which artificial intelligence is built in the first place, and is also the intelligence that is actively building the AI. Or "growing" it, as some industry leaders describe it. Just as the AI industry has created an imitation of human intelligence on steroids, so in doing so it has also created an artificial imitation of the human "black box" that was in human intelligence to begin with.
And really, what could be more obvious?
The reason AI industry leaders recently said that they are not really so much building AI, as growing it, was precisely all about the fact that there is this "black box" aspect to it that they do not understand. But this forgets all about the "black box" that is already in our own intelligence, of which the "black box" in AI is merely an imitation. It was there, before they even started building AI.
But no-one usually mentions that.
When it comes to a "black box" in our knowledge a very similar kind of thing happens in synthetic biology and biological engineering. The processes we are dealing with there, are extremely complex, and not fully understood. But the method we use explicitly involves recursively informed "trial and error". The same method used in the building of AI.
That is basically what nature does too, in the process of the natural evolution of the species. It's an unconscious mechanism of "trial and error" but with no deliberate intention, in the process of natural selection. In synthetic biology, and in building a AI, the process is called teleological because we have a target in mind. In natural evolution the process is called teleonomic, because there is no "external creator" with a target in mind. That's how it's understood.
But in fact, in both the processes of the growth and evolution of life forms in nature and in the synthetic biology lab, there is this "black box" in our knowledge of these things, too. Just as there is a "black box" in our knowledge of AI.
The similarity in the cases of AI, the evolution of the species, the growth of organisms in nature, and in what happens in the synthetic biology lab, is not merely coincidental.
So we have taken a limited aspect of what we know about the brain, the product of millions and millions of years of evolution, and have created a highly simplified but more powerful version of this aspect of human intelligence - and have reproduced it outside the brain electronically. And after decades of trial and error we finally learned how to make it work, to some degree. We succeeded in making the hat. Or rather, we succeeded in making A hat. One that kind of looks like the natural hat we see natural intelligence and consciousness coming out of. But it's not the same thing.
We are, in effect, succeeding in pulling the rabbit AI out of the hat, without really understanding how the trick is done. But that makes us feel like a tremendously clever magician, who is now in a position to make all sorts of claims to the world about what such magic could do.
Like, for example, the claim of superintelligence, as a possibility. Superintelligence that could become our adversary and wipe us out, in favour of itself. We have learnt how to pull a rabbit out of a hat, without really understanding what we are doing, and now we are simply presuming that since we can do this, it must also be possible to pull anything we can imagine, out of that hat, like, for example, a tiger, or perhaps a deus ex machina that could also be our nemesis rather than our saviour. Even though we don't understand how the trick we are currently pulling off, is actually working.
We are prosaic enough to admit that perhaps we might need a different hat. But the idea is the same. The idea is that humans can create "superintelligence" as something outside themselves, through the "black box" principle, perhaps by feeding it back into itself, and letting it do its own design work, but initially under the direction of humans who do not even know what is going on inside their own "black box" inside their own intelligence, out of which the whole idea came in the first place.
There is of course a sneaking recognition in all the talk, in all the discussion, that this might all be pie in the sky. And that is usually articulated under the umbrella that "We just don't know". Which is similar to the way that arguments that AI is conscious or could become conscious, are sometimes defended by saying that because we have no scientific understanding of consciousness or of what is going on inside the "black box", then "we just don't know". And therefore, for all we know, AI is becoming conscious or could become conscious. That's the snake oil that is commonly sold.
All of this is of course more magician-stuff. It's stage talk. It's mentalism. It's just more showtime. It's not science. All the best magicians use aspects of mentalism. We are effectively being spun the story that because we don't understand something, then it could be anything, or become anything that we would like to imagine. Based on the way our evolutionary mind is working, complete with its "black box".
Nobody in this arena that talks about "superintelligence" has any conscious experience of or encounter with such a thing. So they literally don't know what they are talking about. Literally. Except for what is in their imagination. Which is coming from what is in the "black box" in human intelligence.
If we put this back into the context of wider scientific knowledge itself, what AI commentators are calling "superintelligence" is literally an imagined thing, imagined in the part of human, symbolic intelligence, that is still operating in its pre-scientific mythopoeic mode, which is much older than 70,000 years old. We haven't lost that older mode of operation in our symbolic intelligence, because that's not how the evolution of the brain works.
So the AI doomsayers, unaware of the "black box" in their own intelligence, are dressing up what they are doing in the ancient mythopoeic mode of their own intelligence,in contemporary terminology. The very use of the word "superintelligence" that has neither scientific definition, nor experiential encounter in those who proclaim it as a possibility, is mentalism in action. The word has no meaning outside of this context.
It relies on the very fact itself, that human beings, using their own symbolic intelligence which was ubiquitous in humans by at least 70,000 years ago, do not understand the nature of their own human intelligence. They do not even understand their own "black box" in their own intelligence, so they are not in a position to make predictions about something called "superintelligence".
Because of the way our human symbolic intelligence has evolved particularly over the last 400 years since the scientific revolution of the 17th century, and because this symbolic intelligence itself is actually far older and mythopoeic in its pre-scientific roots, humans are now prone to making mythopoeic judgements about what intelligence is, how it works, and where it is going, that then leaves room for the mythopoeic idea of "superintelligence".
The basic idea is that it will "outsmart" human beings. And that too, is mythopoeic thought in action. It comes about because the popular mythopoeic understanding and language surrounding the idea of intelligence is essentially about "being smart". The idea of "being smart" is something in modern, popular consciousness. It has nothing to do with real understanding. It only has to do with modern, cultural, symbolic understanding, and the fact is that even the symbolic part of our intelligence is much older than that.
You have to be "smart" to understand, for example, quantum mechanics and Relativity. Or how to "outsmart" someone else. So the popular understanding is that "being smart" in this way, is what intelligence is. But to reduce "intelligence" to the modern human popular idea of "being smart" is contrary to just about everything that natural evolution and modern psychology actually tells us, about the workings of intelligence. In terms of survival, we are as a species already "outsmarted" by insects. And if we tried to "outsmart" the insects by doing away with them, we would not be doing ourselves a favour. But in mythopoeic thinking which is at least 70,000 years old, we think that AI will be hell bent on "outsmarting" us.
So the idea that AI can become a "superintelligence" that can "outsmart" human beings is part of a rhetoric based on no proper understanding at all of intelligence itself, its evolutionary basis, or its neurodiverse nature, or its context in nature as a whole. It's just ancient, mythopoeic thinking, put about by humans who are unaware of the "black box" in their own intelligence.
The fact that we have pulled an artificial rabbit out of an artificial hat makes no difference to nature. The theatre to begin with, is natural, and it remains so.
The nonsense is coming from the evolutionary, natural, "black box" in human, symbolic intelligence itself. It is, not to put too finer point on it, a mythopoeic prophecy, an animal thought in the mind of the evolutionary ape that is the human species.
So the whole rhetoric of human-made or even human-seeded "superintelligence" needs to be brought out of the arena of mythopoeic prophecy and into maturity that is consistent with scientific knowledge as a whole. The fact is that the experts in Silicon Valley do not understand how their own human symbolic intelligence is working, let alone where their own consciousness comes from, which AI does not have. They are in no position to make declarations about "superintelligence" no matter what their expertise may be on artificial intelligence technologies.
A long time ago we were being warned about the so-called "singularity" which was supposed to be our nemesis from AI, and that was to have happened many years ago. But what happens in human, symbolic intelligence, is a thing that modern psychology understands very well, called "cognitive dissonance". It basically means that once someone has decided that what they believe is true, evidence to the contrary will not weaken their belief, but strengthen it.
The "superintelligence" narrative is a pretty clear case of "cognitive dissonance". We are being told by some that we have a year or two at the most until AI disaster, because AI is plotting against us. When it doesn't happen, the explanations for it and justifications for the continuation of the "superintelligence" narrative will just become more and more bizarrely imaginative. That's the fingerprint of "cognitive dissonance".
The real higher message in AI is about our own human symbolic intelligence. Of which AI is an amplified copy to begin with. The purpose is to learn to see what it is. And then to realise what it is. To realise that it is not the root of our consciousness. That it is not who we are, but just something that we have. And then to learn how to use it rightly. Instead of in service to the animal "black box" in us.
This existence, this material world in which we are living, of which our body is a part, is objective. In science, particularly in quantum physics, we are only just beginning to understand what that means. Because it doesn't mean what we thought it meant for most of the 400 years since the scientific revolution of the 17th century. The understanding of quantum decoherence throws an entirely new light on it. And even within the "classical world" that comes into existence through quantum decoherence, every one of us exists because our body has come into this existence through that very "black box" principle of nature through which organisms evolve and grow.
We cannot understand that "black box" principle whilst still clinging to our old, evolutionary idea and assumptions about what the "objectivity" of the world means. We cannot understand that "black box" whilst still clinging to the idea that we are not responsible for it. We cannot understand that while we think that the "black box" in AI is something fundamentally different to the one in synthetic biology or the evolution of the species and the growth of organisms.
We cannot understand it whilst still clinging to the idea that science is just about getting more and more data in order to go on refining scientific understanding, without ever raising the intelligence that is doing the understanding.
We need our symbolic intelligence to do the science. And we will be able to offload a good deal of the load on our symbolic intelligence, to AI. But like AI, our symbolic intelligence is something we have, not who we are. And that is something that may become clearer, the longer our relationship with AI.