The video associated with this piece can be found here: https://youtu.be/lrjY4fR-qMU
I have written before about the possible links between free will, consciousness, creativity, explanations, knowledge and choice here: https://www.bretthall.org/free-will-consciousness-creativity-explanations-knowledge-and-choice.html That article was itself me riffing off a brief passage in “The Beginning of Infinity” by David Deutsch which appears on page 157 which is in Chapter 7 - artificial creativity. David writes (and here by “AI” he is using the earlier traditional usage - we would now say AGI)
“AI abilities must have some sort of universality: special-purpose thinking would not count as thinking in the sense Turing intended. My guess is that every AI is a person: a general-purpose explainer. It is conceivable that there are other levels of universality between AI and ‘universal explainer/constructor’, and perhaps separate levels for those associated attributes like consciousness. But those attributes all seem to have arrived in one jump to universality in humans, and, although we have little explanation of any of them, I know of no plausible argument that they are at different levels or can be achieved independently of each other. So I tentatively assume that they cannot.”
We have little explanation of any of them - and nor do I nor do I pretend to here and now. But I do want to probe them a little further and here I want to focus more narrowly today just on consciousness and creativity.
Have you ever done something on “autopilot”? Many of us have had the experience whether we walk, drive or ride a regular route between work and home or the shops and home or wherever…and home that we can make the journey on a particular day and never quite recall exactly how it is we got there.
Others know what it’s like to work in some place assembling a thing and doing the same repetitive motion over and again. We have a term “second nature” to describe those activities so familiar to us they appear to be a part of our being - a natural part. It’s second nature to me how to use parts of video editing software to make my podcasts these days. I used to have to think carefully about each and every step - but no more. Second nature - or doing things on autopilot - tends to apply to activities that become rather straight forward to do and which we are not attempting to improve. So sport in general does not count because although we might be highly competent at a particular skill there, we are also trying hard to improve if only in the slightest way possible. And we are having fun with that exact thing. It is more often the mundane things that can tend to be handed over to the unconscious or at least less conscious part of the mind.
Assembling widgets, knitting, driving a car or riding a bike along a familiar road is like this - those things can often be largely automated by our own minds and bodies. We may become not so much bored by such an activity but we are not actively trying to improve our performance in that activity. We reach a level of competence, we know it “gets the job done” and at that point it becomes so routine we can actually begin to allow our mind to do other things.
We were doing that activity automatically - or rather unconsciously. It is not to say we were unconscious while doing it, but whatever part of the mind is devoted to controlling that part of our behaviour is somehow operating without our consciously paying attention to it. Or at least much attention to it. Creativity may still be going on, but it is of an unconscious kind. It frees us up to think of other things - to put most of our creativity into paying attention to our other conscious thoughts.
I think this capacity of human beings to “do things on autopilot” or “unconsciously” (even though it is not strictly unconscious) is a clue as to the connection between consciousness and creativity.
If I am riding an unfamiliar route on my bike I can be listening intently to an audiobook and might travel many kilometres and by the end of the journey have a reasonably good recollection of each part of that journey. I can recall landmarks, a particular dog that barked in some yard, a child running in the opposite direction pursued by their mother pushing a pram. I have a more vivid recollection if this is the first time I have been along that particular route. However, in the case of another far more familiar route, one I have completed hundreds of times before, the activity itself manages to get done by a part of my mind not paying attention to the landmarks. Indeed I can get all the way home and catch myself and wonder “how did I get here”. This is especially true if engrossed in something else (for those concerned about safety, this is on a track devoted entirely to pedestrians and bicycles and I ride only slowly).
This phenomenon by the way has a name: “highway hypnosis” or “road hypnosis” https://en.wikipedia.org/wiki/Highway_hypnosis and can be very dangerous. I do not wish to discuss the dangers and will rather just refer to the somewhat less dangerous form I am personally more familiar with - the riding of a bicycle and not on any road (much less a highway) alongside many other vehicles travelling at high speeds, carrying passengers.
The analogue of “highway hypnosis” crops up in many places. Riding a bike along a familiar track is obviously very closely related. The riding of the bike becomes almost like breathing. It is not as if I cannot suddenly take control of the autonomic nervous system in this regard. Should a child dart across the path suddenly I can be brought back into the present moment, so to speak, and slam on the brakes and there we have it: creativity is back in the saddle and the audiobook I was listening to entirely forgotten as I try to avoid injury to us both. One might even apply consciousness directly to the breath at that point in an attempt to calm down. Take a deep breath, and all that. Consciousness and creativity are focussed on the here and now. But if I resume the journey after some time I can return to paying attention to just what I was listening to - but I have to rewind to pick up what I have missed. Consciousness is a fickle thing that does not like to be split between activities. But it can readily switch in this case to, once more, not paying attention - certainly not in the same way - to riding, steering, breaking and accelerating. Instead I am focussed back on my audio. Having done this route hundreds of times before the instructions for this journey are there in memory somehow connected to behaviour while bypassing consciousness - while bypassing creativity. But clearly not entirely. It’s not like I am entirely unconscious of the journey. Highway hypnosis is an altered state of consciousness. Not a truly unconscious state. But as a matter of subjectivity, it seems probative of this consciousness and creativity link.
What makes the difference between a truly “dumb” unconscious computer (or say a robot) which is able to move from A to B reliably and a person doing so in the way I described “unconsciously” (or under highway hypnosis) is that of course the person never is unconscious. To describe their activity as being completed “unconsciously” is a manner of speaking. But it is not entirely meaningless. They are of course conscious, but the part of the mind they are primarily aware of and which is laying down memories is the conscious part while the part actually completing “the task” of getting from A to B has been handed over to a part of the mind that is simply performing some instructions with little, perhaps no, creativity and seems not to be recorded in memory. In that sense and to that extent the person is emulating the behaviour of an AI. Completing the task unconsciously while doing other things with their consciousness (like thinking of other more interesting things - working through problems).
We of course do not know how the mind works in any fine grained way - this is just a corollary of David’s “we have little explanation of any of them” - the features minds have. But we do know as a matter of personal subjectivity that there are intrusions possible from the conscious part of the mind into the unconscious and vice versa as I have already indicated. But the idea people can learn to perform routine tasks in a way that is so reliable they no longer need to pay barely any attention to them is a clue about the link between consciousness and creativity. Certainly anything unconscious is not creative. A robot or any existing computer is not conscious. Assume some robot is. Which part of it? Not its hands for it has no sensation. Its eyes? But those are just cameras. The processor? That’s just hardware. A person’s mind is conscious and we experience qualia - sensations and we experience sensation precisely in those places where we have nerves for sensation. That aside we know robots are not creative. If the unconscious is not creative, does the converse hold as well? Is the uncreative always unconscious? We cannot say and this is controversial. I have written before about whether it follows if other animals, being uncreative are conscious in any way that resembles what is going on with people. I will not recapitulate that argument here and now - the link is at https://www.bretthall.org/humans-and-other-animals.html
Can we learn anything about this link between consciousness and creativity ourselves as a matter of deliberate introspection? When some people meditate they report a sense of “not having a self” - but if that sense is just a sense that one is not being creative at that moment (because of some deliberate intention not to be so - to deliberately “drop one’s problem” so to speak) it is unsurprising to conclude “there is no self” if the self is a creative thing that deals with problems. If you look and detect no creativity and the self is “that which is creative” then it would seem to be a reasonable conclusion to draw. Of course the self merely has the potential to be creative. So if it is not being actualised because it is in some kind of idle state it should be unsurprising not to notice it. Or rather, to notice it not doing anything and for some then “so it’s not really there at all” seems a reasonable conclusion to draw.
Deliberately not being creative (not being “lost in thought”) would then be the perfect way to be aware of not being a self. Of course this is all a kind of empiricist debate: the idea that because we can or cannot sense something about ourselves is no clue whatsoever that such a thing exists (or does not exist). The inability to have a sense of self is neither proof nor refutation of anything about the self. It is, at best, a demonstration of one’s fallibility - even about the fact of consciousness. What the self is, is a question in need of a good explanation - not merely a sensation or an experience. Insofar as it is an experience once again we are retuned to the question: what is the explanation of that experience, or experience, period.
I have no answers here about precisely the link between consciousness and creativity. If we pay attention, which is to say “direct our consciousness” to a particular task what we are doing is bringing into our awareness some activity so that we can think about it - play with it, problem solve and be creative with it. Other activities we are familiar with very much can be done, loosely speaking, unconsciously and although we do not mean “strictly unconsciously” we do mean something like “automatically” - like an automaton. We are not quite problem solving in the same way to the same degree when we are on autopilot driving home or riding a familiar path or performing that same daily task.
Or in other words in that particular domain we are not learning - truly learning. We are entities that learn, we cannot switch that capacity off, but we do seem to have the capacity to learn how to learn while doing other things we have already learned…if that is not needlessly convoluted. To rephrase: once we have conjectured the knowledge then tested it in the world and it solves our problem then this knowledge can run like an algorithm it seems directing our behaviour while that which we really are - the conscious creative entity that is a person is able to move on to learning other things. This happens when we are on so-called autopilot.
Our creativity and consciousness is elsewhere for a time working on problems for which we do not know solutions just there (like, for example: what might be said next in some audiobook). But the unconscious mind - that under “highway hypnosis” while it might be “solving problems” in some sense (knowing which way to steer and when to peddle a little harder or when to apply the brake) cannot deal with the unexpected quite so well. For if it does encounter the problem not encountered before, consulting its library of instructions about “what to do next” is of no use and so a message needs to be sent quick smart to take over consciousness again. Why? Because we need some creative thinking.
As I have flagged, I think these facts are important clues about not merely a link, but a necessary connection if not a continuum of a kind between consciousness and creativity and is one reason my guess is that it just is not possible in the AGI debate to think we could ever build a system that is truly creative but simultaneously unconscious. I think the idea that is possible arises from an empiricist mistake people make when thinking deeply about, for example, what happens when they have been meditating. In thinking back to their experience they feel they have had a sensation of what consciousness is in some direct way and conclude (wrongly) that it can be entirely divorced from thinking because they feel as if they are not identical to thoughts, while only having them. This may contain a grain of truth: one is not identical to their thoughts (they merely have thoughts and ideas as I have argued elsewhere: https://www.bretthall.org/critically-creative-3.html )
But if consciousness is something more like “the potential for creativity” then we can witness rare occasions such as during highway hypnosis, flow states and mindfulness where our creativity is not being utilised hardly at all during that activity and is freed up to do otherwise. But the point is: it - creativity - is there doing other work and consciousness is there aware of that other work while neither creativity nor consciousness are devoted to the mundane routine task.
We ourselves experience our peak of creativity and productivity when most conscious - most awake and alive precisely in those places we are devoting the most of our attention to. When tired or working on autopilot, we are less aware of what we are doing. Less conscious in a sense and necessarily I would say less creative about those things.
I do not think this is merely a correlation but something deeper. Is creativity consciousness exactly? Are they two words labelling some unifying quality of a mind which a future theory thereof will illuminate for us? Are both but two facets of some third more complicated thing which incorporates equally contentious notions of free will, qualia and more besides? We don’t know.
But it should be telling that our own experience of consciousness and of being creative is telling us something about reality even if it is not easy to put these experiences into words. But the fact we cannot easily speak of these things is no reason to think we cannot speak of these things. It just means we do not know how yet. We lack explanations. But that’s just to say: we have some problems. And until these problems are solved, attempting to provide good explanations of whether some AI is or is not conscious or creative will be entirely elusive if we cannot provide an equally good explanation of how it is that we humans are.
I have written before about the possible links between free will, consciousness, creativity, explanations, knowledge and choice here: https://www.bretthall.org/free-will-consciousness-creativity-explanations-knowledge-and-choice.html That article was itself me riffing off a brief passage in “The Beginning of Infinity” by David Deutsch which appears on page 157 which is in Chapter 7 - artificial creativity. David writes (and here by “AI” he is using the earlier traditional usage - we would now say AGI)
“AI abilities must have some sort of universality: special-purpose thinking would not count as thinking in the sense Turing intended. My guess is that every AI is a person: a general-purpose explainer. It is conceivable that there are other levels of universality between AI and ‘universal explainer/constructor’, and perhaps separate levels for those associated attributes like consciousness. But those attributes all seem to have arrived in one jump to universality in humans, and, although we have little explanation of any of them, I know of no plausible argument that they are at different levels or can be achieved independently of each other. So I tentatively assume that they cannot.”
We have little explanation of any of them - and nor do I nor do I pretend to here and now. But I do want to probe them a little further and here I want to focus more narrowly today just on consciousness and creativity.
Have you ever done something on “autopilot”? Many of us have had the experience whether we walk, drive or ride a regular route between work and home or the shops and home or wherever…and home that we can make the journey on a particular day and never quite recall exactly how it is we got there.
Others know what it’s like to work in some place assembling a thing and doing the same repetitive motion over and again. We have a term “second nature” to describe those activities so familiar to us they appear to be a part of our being - a natural part. It’s second nature to me how to use parts of video editing software to make my podcasts these days. I used to have to think carefully about each and every step - but no more. Second nature - or doing things on autopilot - tends to apply to activities that become rather straight forward to do and which we are not attempting to improve. So sport in general does not count because although we might be highly competent at a particular skill there, we are also trying hard to improve if only in the slightest way possible. And we are having fun with that exact thing. It is more often the mundane things that can tend to be handed over to the unconscious or at least less conscious part of the mind.
Assembling widgets, knitting, driving a car or riding a bike along a familiar road is like this - those things can often be largely automated by our own minds and bodies. We may become not so much bored by such an activity but we are not actively trying to improve our performance in that activity. We reach a level of competence, we know it “gets the job done” and at that point it becomes so routine we can actually begin to allow our mind to do other things.
We were doing that activity automatically - or rather unconsciously. It is not to say we were unconscious while doing it, but whatever part of the mind is devoted to controlling that part of our behaviour is somehow operating without our consciously paying attention to it. Or at least much attention to it. Creativity may still be going on, but it is of an unconscious kind. It frees us up to think of other things - to put most of our creativity into paying attention to our other conscious thoughts.
I think this capacity of human beings to “do things on autopilot” or “unconsciously” (even though it is not strictly unconscious) is a clue as to the connection between consciousness and creativity.
If I am riding an unfamiliar route on my bike I can be listening intently to an audiobook and might travel many kilometres and by the end of the journey have a reasonably good recollection of each part of that journey. I can recall landmarks, a particular dog that barked in some yard, a child running in the opposite direction pursued by their mother pushing a pram. I have a more vivid recollection if this is the first time I have been along that particular route. However, in the case of another far more familiar route, one I have completed hundreds of times before, the activity itself manages to get done by a part of my mind not paying attention to the landmarks. Indeed I can get all the way home and catch myself and wonder “how did I get here”. This is especially true if engrossed in something else (for those concerned about safety, this is on a track devoted entirely to pedestrians and bicycles and I ride only slowly).
This phenomenon by the way has a name: “highway hypnosis” or “road hypnosis” https://en.wikipedia.org/wiki/Highway_hypnosis and can be very dangerous. I do not wish to discuss the dangers and will rather just refer to the somewhat less dangerous form I am personally more familiar with - the riding of a bicycle and not on any road (much less a highway) alongside many other vehicles travelling at high speeds, carrying passengers.
The analogue of “highway hypnosis” crops up in many places. Riding a bike along a familiar track is obviously very closely related. The riding of the bike becomes almost like breathing. It is not as if I cannot suddenly take control of the autonomic nervous system in this regard. Should a child dart across the path suddenly I can be brought back into the present moment, so to speak, and slam on the brakes and there we have it: creativity is back in the saddle and the audiobook I was listening to entirely forgotten as I try to avoid injury to us both. One might even apply consciousness directly to the breath at that point in an attempt to calm down. Take a deep breath, and all that. Consciousness and creativity are focussed on the here and now. But if I resume the journey after some time I can return to paying attention to just what I was listening to - but I have to rewind to pick up what I have missed. Consciousness is a fickle thing that does not like to be split between activities. But it can readily switch in this case to, once more, not paying attention - certainly not in the same way - to riding, steering, breaking and accelerating. Instead I am focussed back on my audio. Having done this route hundreds of times before the instructions for this journey are there in memory somehow connected to behaviour while bypassing consciousness - while bypassing creativity. But clearly not entirely. It’s not like I am entirely unconscious of the journey. Highway hypnosis is an altered state of consciousness. Not a truly unconscious state. But as a matter of subjectivity, it seems probative of this consciousness and creativity link.
What makes the difference between a truly “dumb” unconscious computer (or say a robot) which is able to move from A to B reliably and a person doing so in the way I described “unconsciously” (or under highway hypnosis) is that of course the person never is unconscious. To describe their activity as being completed “unconsciously” is a manner of speaking. But it is not entirely meaningless. They are of course conscious, but the part of the mind they are primarily aware of and which is laying down memories is the conscious part while the part actually completing “the task” of getting from A to B has been handed over to a part of the mind that is simply performing some instructions with little, perhaps no, creativity and seems not to be recorded in memory. In that sense and to that extent the person is emulating the behaviour of an AI. Completing the task unconsciously while doing other things with their consciousness (like thinking of other more interesting things - working through problems).
We of course do not know how the mind works in any fine grained way - this is just a corollary of David’s “we have little explanation of any of them” - the features minds have. But we do know as a matter of personal subjectivity that there are intrusions possible from the conscious part of the mind into the unconscious and vice versa as I have already indicated. But the idea people can learn to perform routine tasks in a way that is so reliable they no longer need to pay barely any attention to them is a clue about the link between consciousness and creativity. Certainly anything unconscious is not creative. A robot or any existing computer is not conscious. Assume some robot is. Which part of it? Not its hands for it has no sensation. Its eyes? But those are just cameras. The processor? That’s just hardware. A person’s mind is conscious and we experience qualia - sensations and we experience sensation precisely in those places where we have nerves for sensation. That aside we know robots are not creative. If the unconscious is not creative, does the converse hold as well? Is the uncreative always unconscious? We cannot say and this is controversial. I have written before about whether it follows if other animals, being uncreative are conscious in any way that resembles what is going on with people. I will not recapitulate that argument here and now - the link is at https://www.bretthall.org/humans-and-other-animals.html
Can we learn anything about this link between consciousness and creativity ourselves as a matter of deliberate introspection? When some people meditate they report a sense of “not having a self” - but if that sense is just a sense that one is not being creative at that moment (because of some deliberate intention not to be so - to deliberately “drop one’s problem” so to speak) it is unsurprising to conclude “there is no self” if the self is a creative thing that deals with problems. If you look and detect no creativity and the self is “that which is creative” then it would seem to be a reasonable conclusion to draw. Of course the self merely has the potential to be creative. So if it is not being actualised because it is in some kind of idle state it should be unsurprising not to notice it. Or rather, to notice it not doing anything and for some then “so it’s not really there at all” seems a reasonable conclusion to draw.
Deliberately not being creative (not being “lost in thought”) would then be the perfect way to be aware of not being a self. Of course this is all a kind of empiricist debate: the idea that because we can or cannot sense something about ourselves is no clue whatsoever that such a thing exists (or does not exist). The inability to have a sense of self is neither proof nor refutation of anything about the self. It is, at best, a demonstration of one’s fallibility - even about the fact of consciousness. What the self is, is a question in need of a good explanation - not merely a sensation or an experience. Insofar as it is an experience once again we are retuned to the question: what is the explanation of that experience, or experience, period.
I have no answers here about precisely the link between consciousness and creativity. If we pay attention, which is to say “direct our consciousness” to a particular task what we are doing is bringing into our awareness some activity so that we can think about it - play with it, problem solve and be creative with it. Other activities we are familiar with very much can be done, loosely speaking, unconsciously and although we do not mean “strictly unconsciously” we do mean something like “automatically” - like an automaton. We are not quite problem solving in the same way to the same degree when we are on autopilot driving home or riding a familiar path or performing that same daily task.
Or in other words in that particular domain we are not learning - truly learning. We are entities that learn, we cannot switch that capacity off, but we do seem to have the capacity to learn how to learn while doing other things we have already learned…if that is not needlessly convoluted. To rephrase: once we have conjectured the knowledge then tested it in the world and it solves our problem then this knowledge can run like an algorithm it seems directing our behaviour while that which we really are - the conscious creative entity that is a person is able to move on to learning other things. This happens when we are on so-called autopilot.
Our creativity and consciousness is elsewhere for a time working on problems for which we do not know solutions just there (like, for example: what might be said next in some audiobook). But the unconscious mind - that under “highway hypnosis” while it might be “solving problems” in some sense (knowing which way to steer and when to peddle a little harder or when to apply the brake) cannot deal with the unexpected quite so well. For if it does encounter the problem not encountered before, consulting its library of instructions about “what to do next” is of no use and so a message needs to be sent quick smart to take over consciousness again. Why? Because we need some creative thinking.
As I have flagged, I think these facts are important clues about not merely a link, but a necessary connection if not a continuum of a kind between consciousness and creativity and is one reason my guess is that it just is not possible in the AGI debate to think we could ever build a system that is truly creative but simultaneously unconscious. I think the idea that is possible arises from an empiricist mistake people make when thinking deeply about, for example, what happens when they have been meditating. In thinking back to their experience they feel they have had a sensation of what consciousness is in some direct way and conclude (wrongly) that it can be entirely divorced from thinking because they feel as if they are not identical to thoughts, while only having them. This may contain a grain of truth: one is not identical to their thoughts (they merely have thoughts and ideas as I have argued elsewhere: https://www.bretthall.org/critically-creative-3.html )
But if consciousness is something more like “the potential for creativity” then we can witness rare occasions such as during highway hypnosis, flow states and mindfulness where our creativity is not being utilised hardly at all during that activity and is freed up to do otherwise. But the point is: it - creativity - is there doing other work and consciousness is there aware of that other work while neither creativity nor consciousness are devoted to the mundane routine task.
We ourselves experience our peak of creativity and productivity when most conscious - most awake and alive precisely in those places we are devoting the most of our attention to. When tired or working on autopilot, we are less aware of what we are doing. Less conscious in a sense and necessarily I would say less creative about those things.
I do not think this is merely a correlation but something deeper. Is creativity consciousness exactly? Are they two words labelling some unifying quality of a mind which a future theory thereof will illuminate for us? Are both but two facets of some third more complicated thing which incorporates equally contentious notions of free will, qualia and more besides? We don’t know.
But it should be telling that our own experience of consciousness and of being creative is telling us something about reality even if it is not easy to put these experiences into words. But the fact we cannot easily speak of these things is no reason to think we cannot speak of these things. It just means we do not know how yet. We lack explanations. But that’s just to say: we have some problems. And until these problems are solved, attempting to provide good explanations of whether some AI is or is not conscious or creative will be entirely elusive if we cannot provide an equally good explanation of how it is that we humans are.