How does Human Intelligence differ from Artificial Intelligence (AI)?

How does human intelligence differ from artificial intelligence?

 

AI is here among us, and increasingly so, in every aspect of our lives.  It is accumulating all the data we humans gather and publish, and presenting itself as an authority.  It speaks our language, answers our questions, explains and solves our puzzles, and performs difficult mental tasks with lightening speed.  AI faces are appearing on our screens, and embodied AIs are appearing in robot form, clothed in ever more sophisticated humanoid bodies which are increasingly able to carry out our physical tasks.  They are coming to resemble us more and more in their verbal expression, their screen image, their robot emobodiment - only they have immediate access to much more data than us,

and a far greater power to manipulate mass data at speed, and they are built of harder stuff than us and have greater mechanical strength, they are unemcombered with feelings, and available to do our bidding.   

We Robot

Elon Musk recently presented the Tesla robot Optumus 5'8" launched in an orchestrated extravaganza entitled 'We Robot' (after the Sci-Fi author Isaac Azomov's famous book 'I robot' ), where he marvelled,  'The Optimus robots will walk among you!'.  He tells us this robot can "...basically do anything you want.  It can be a teacher, babysit your kids, walk your dog, mow your lawn, get the groceries, just be your friend, serve drinks.  Whatever you can think of, it can do."  

For the time being humanity is, it seems, largely content to use AI's, to make money from them, to let them get involved in our lives in verbal and physical form and do our bidding without asking too many questions.  AIs are employed as tools without muchthought to we humans who create it, the bearers of the very human intelligence that AI is an artificial version of.  Of those who do think about AI, more and more are asking, is Artificial Intelligence essentiallly artificial, or does it at least have the potential to be much the same as human intelligence?  Are human brains essentially 'organic evolved computers' as the philosopher Daniel Dennett put it, are we humans in reality

mechanical, wondrously complicated robots?  Going further, some of us ask, is AI potentially better than human intelligence?  Are machines better at learning than humans?  At some point, could AI be considered conscious?  At some point, with access to all the information we give it to enable it to control us, will it take over the world, and run it rather better than we do, considering the mess we humans are making of everything? (the illustration shows the robot Atlas, by Boston Dynamics).

 

 

We humans

To answer these kinds of philosophical and increasingly pressing world questions, we need a clear understanding not only of artificial intelligence and how it works, but of our human intelligence, which Artificial Intelligence imitates, our human learning and consciousness.  We need to know and understand our own human thinking, the very force which conceived of AI and allowed it to be brought into the world.   Only then will we be in a position to see how AI differs from our intelligence in an essential way.  We need to start from our own actuall human learning, before moving on to machine learning.  

A proposed basic understanding of human learning

Looking calmly and carefully at how I function, how I proceed through my conscious life, a world appears before me, and relates to an inner world, which is intimate, and within which I have my own experiences.  The inner world I call ‘me’ or ‘I’, the outer experience which appears before me I call ‘the world’.  

I have some influence on how I relate to this world which appears to me as outside.  I can know about it, and get to know it, so that I can act on it and within it in a more informed way.  I can also observe and know about my inner world, what kind of reactions and actions take place in here, and I have the possibility to change the way I think, and my new thought-out view can change the way I feel, and in turn change the way I observe and think further...and so on.  This possibility applies to my learning relationship with the outer world too.  I can seek to know and change in all areas of my outer experience, and for my inner being, I can take a path of self-development towards my true self - I can try to find out who is in here.  I can ask “who am I being at the moment?”.  I can ask ‘who am I?”.

This contrasting experience of self and world is the basis of my consciousness. If I want to know and understand anything about the outer world which appears outside me, or my own inner world, I start by observing in some way or other, and by thinking what I observe, at least to some extent - it might be more or less conscious, and will vary according to my circumstances and ability.  

I have named a contrast, a polarity, I have identified and verified for myself, within my

own experience, two worlds, an inner and an outer.  However, if I really slow things down and observe carefully what I am doing when I observe an object, for example, I can discover that my thinking penetrates my observing.  In reality there is no separate world out there, which I think about, which I add thoughts to, nicely separated inside myself.  There is an interraction of some sort between 'me', observations and thoughts arising from a thinking process.

Expermiment to observe how thinking penetrates our experience of the observed world

• Take an object which you are unfamiliar with, which you don't know the name or purpose of
• Observe and describe the object
• Observe and describe how you observe it, at the level of your thinking, feeling and willing
• Once you know what the object is, compare your processes with the above

In reality as a conscious human I think the world that appears before me and the world that appears before me is penetrated with my thinking,  so I become aware of it in the form of my thinking consciousness.  So that ‘I’  become aware of it in the form of my thinking consciousness.  I’m creating a knowing-I all the way through this experience as much as I am creating 'knowledge' -  that is to say, I am learning. My level of thinking consciousness can vary at any time of my life or in any moment, depending on my ability, knowledge, situation and so on,  but to some extent my consciousness is an observation-thinking experience.

When it comes to observing and thinking myself, I can observe myself in an outer way, for example, through my actions and their effects in the observable world, by becoming aware of how others see me, by observing the course of my life, or by reflecting on the unfolding of the day I have just lived.  I can also observe my inner world from the inside.  I do not see this inner world with my physical eyes.  How do I become aware of it, that is observe and think my inner world?  I am going to try this out now, as I write these words.  As I turn my attention to this task I stop my outer activity, I close my eyes and grow still and quiet in my outer observing-thinking, my intention is to observe my inner world from the inside:  I ask myself 'what is happening in here?'.  My physical eyes are closed, but my inner eye is open to receiving information, there is an opening, a pure 'observing' type of conscious attention, actively receptive...I can ask myself, 'what is happening in my head?' for example, and observe.  I can ask myself 'what am I feeling?'  or 'what do I want and want to do?  I can also pick up information about my physical state in this way, by inquiring into any areas of pain or discomfort, or healthy and happy states.  I can 'roam around' inside and experience something in response to my thinking questioning and afterwards to some extent at least, express this observation in words.

Observation and thinking are a prerequisite for, a necessary door to all conscious learning experience. 

Are we always learning?

As I explore this topic I asked myself, 'Are we always learning?'.  I tested this out by quietening myself, gathering my observations, and thinking;   I answer ‘yes’ we are always learning, because each time I open my eyes or any of my senses to the outer world something new appears, even if it is the ‘same’ object that I saw yesterday, or a few moments before, something will have changed for me, perhaps nothing more than a subtle change of light, of angle of view, accompanied by a subtle development in me over time, in the inner place which I am coming from at that moment - and so I am always thinking new observations and learning.

I can be in a situation where I react to what appears in the world, but I also have the possibility set my own goals and try to achieve them.  My goals are in the form of ideas of some sort.  I can observe these ideas and decide I want to act in them.  Here things are turned around.  My ideas, my thinking,  acts on the world to create something, through my willing and bodily being.  I am no longer only reacting to and thinking the given observed world.  I am creating something in the world, which I can observe in the outer world.  I can observe how it appears and plays out in that outer shared world, rather than merely ruminating and speculating within my own sphere of consciousness.

Machine Learning

Artificial Intelligence (AI) experts and those working in the field, when asked what AI is and what machine learning is, tend to give self-contained technical replies.  They understand the nature and limits of AI in a practical way and are clear in this sense on how AI differs from human intelligence.  Here is an example of such a reply to the question 'what is machine learning?':

"Machine learning is a subset of Artificial Intelligence.  Given a set of inputs and their correct outputs, the process of learning is for a system to lean towards producing outputs as close as possible to the correct outputs". 

We can pause to absorb this information, get a feeling for its quality and how is, how it feels when we 'think' it for ourselves.  From inner observation, what kind of thinking am I using when I read, absorb and understand these words?  Using our basic intellectual thinking about this succinct text, we can say, in human terms, that machine-learning is about a machine working towards producing an answer which is already decided upon and judged 'correct', an answer in response to any given question which it connects to and interprets from its own stored data.   'Learning' in this machine context,  is about recognizing the expressed question within data from the past, and matching it to the pre-defined answer - it is about creating automatic pathways from question to answer.  It is about generalising sufficiently to be able to answer any specific question.  From this narrow definition of learning, the question can come up, is it the same for human learning?  Is this the extent and limit of human learning, or are we cabable of something more, something fundamentally different?  Straight away we can see that the machine does not 'see', it is blind, it does not 'observe' directly,  it has to be programmed with coded symbols of human observation.  Nor does it think actively, nor is it aware of a thinking self, nor does it operate from a sense of unified self, it is a set, a bundled accumulation of automatic connections.

Here is a second example of a technical definition of machine learning, this time including a human element:

"Human workers train a network of neuron-equivalents by inputting data, we examine the output, and then update the network of neuron-quivalents so they reach the correct outputs  - in order to correct any errors."

Humans put the data in, humans 'add' to the data of correct responses, and create pathways between the two, until the machine demonstrates more and more 'correct' answers.  The machine learning process requires consistent human intervention to judge whether or not the output is 'right' in the context of the input, and adjust it accordingly, to allow the machines to slough off, bit by bit, any 'errors' as defined by its human creators.  At some point the machine has sufficient data and connective pathways to run by itself, to manipulate any question and its own vast databanks sufficiently to answer correctly without further intervention.  Humans are in effect putting something of themselves into the machine and hoping it will run successfully by itself.

If we ask ChatGPT the question ‘Are you conscious’ it 'replies' (in reality this is an artificial semblance of intelligent reply): 

"No, I’m not conscious.  I don’t have awareness or feelings;  I process and generate text based on patterns in data’. 

Chat GPT has been programmed and 'prompted' to 'tell' us it is not conscious, and to equate consciousness with awareness and feelings, and to define its own task as 'generating text based on data.' 

The French AI 'Le Chat Mistral' gives a more people-friendly answer to the question "What is machine learning?"  

"Machine learning is a type of artificial intelligence (AI) that allows computers to learn from data, rather than being explicitely programmed (that is for each individual question and answer).  It's like teaching a child to recognize a cat.  You don't give the child a set of rules (like, it has pointy ears, whiskers etc) you show the child many pictures of cats and the child learns to recognize what makes a cat a cat. 

How does it work? Imagine you are teaching a computer to recognize cats.  You show it lots of pictures, some with cats, some without.  The computer begins by making guesses about which one is a cat, which one isn't a cat.  Then you tell the computer each time whether it is right or wrong.  Based on that feedback, the computer adjusts its guesses a little bit.  It does this over and over again, each time getting a bit better at discerning cats from non-cats.  That's the basic idea of machine learning!"

Computers of all sorts work uniquely on a binary system, represented by the numbers 0 and 1.  Everything must be reduced to this kind of 'yes or no, right or wrong' format, in this case, cat or not-cat.  The machine receives a 'defined object' in the form of a measure of information, not through direct observation as humans do.   A human observes a cat, the machne however receives its equivalent of 'observation' it in a converted, compacted form, reducible to combinations of 0 and 1.  For example, if the information was originally in words these words will be converted into 'tokens'.  The machine is trained to link the input token to a correspending 'output' token and turn this output token back into written or spoken words in order to give us a 'reply'.   Cat input object' is converted to cat-token, which is linked up to 'cat-output token', and translated to the naming word 'cat'.  The machine robot can can 'speak our language', but its invisible inner processes are worlds apart from ours.

How is the task of teaching the machine to learn carried out?  The machine is shown two sets of objects, cats and non-cats, and must 'learn' to distinguish between those things.  It does this by a statistically significant amount of guessing and being informed each time if it guessed right or wrong,  in the case of cat recognition this would amount to millions of confirmed or corrected guesses, until it amasses enough information to improve its probability of making more and more correct guesses.  It is able to do this from the sheer quantity of information it has and the sheer quantity of correlations it retains.  It never actually sees a cat and knows 'this is a cat', this state that we humans  experience all the time, even as very young children. 

"It's like teaching a child to recognize a cat.  You don't give the child a set of rules (like, it has pointy ears, whiskers etc) you show the child many pictures of cats and the child learns to recognize what makes a cat a cat."

Even very young children do not need millions of correlated experiences to learn to recognize a cat, so something very different is going on with human learning, we are not learning from repeated generalised correlations.  For a child, the process of being taught the word 'cat', in order to name the being behind these multiple experiences, does involve repetition.  A child can have multiple experiences of a cat - involving sight sound touch smell movement and all the information from our senses - and it can meet different types of cats in different circumstances, yet all this is gathered to name one being with the word cat.   The word cat has to be repeatedly spoken and correlated with each experience - however for humans,  'What makes a cat a cat' is not learned by repeatedly being shown different snapshot pictures, it is not something we 'learn' or can be programmed with through mass generalisation.  Somehow we have an experience of meeting and knowing what we call 'cat', in an instant of recognition.  This is how we humans think.  We learn to name concepts with words, the concepts gather together to make more and more unified concepts, for example, I see an individual cat 'Catrina' and many individual cats and images of cats, and this gathers around a unifying concept 'cat'.

We can grasp such technical descriptions of what AI is, at least to some extent,  with our intellectual, academic comprehension, with one aspect of our human intelligence,  but is this really the deep answer to the deep question we are asking when we want to know;  is there an essential difference between human and artificial inteligence ?

Do we humans have access to any other kind of intelligence?  A creative intelligence, for example, or emotional, moral or intuitive intelligence?  Are we rather the bearers, by our nature, of a spiritual activity which operates outside the bounds of the mechanics of our body-brain make-up?  If so, we could say clearly that the human thinking behind our intelligence and learning is a spiritual activity, which gives us access to a spiritual realm, a ‘cosmic intelligence’ which is beyond our bodily make-up and limitations.  By contrast, machines which learn, the‘AI’s, are our creatures:  we created them and they are directly answerable to us, they do not have direct access to our uniquely human, spiritual activity of thinking and its cosmic source.

In order to demonstrate this proposition, that we are the bearers of a spiritual activity which gives us access to a spiritual realm outside our bodily limitations, we have to demonstrate that our human thinking is a spiritual activity.  How can we do this?  We can only demonstrate it by experience - just as we rely on experience of one sort or another as the basis of all our human learning, knowing and intelligence.  If we could manage to experience the spiritual activity of our thinking, then we could as a result of this experience, talk about it.  We could give a basic description of it to someone who had not had this experience, give the basic orientation and co-ordinates for finding it,  map out how to reach it, for another person to follow or at least know about.  Yet, we can only demonstrate it to ourselves by an experience we can observe and think about, by an experience of this fact of our human nature and way of being.  No amount of intellectual theories or speculation can add light to the question, we must ourslves experience this aspect of our true and actual nature.

How then can we experience the spiritual activity of our own human thinking?  This is one of the principle goals of stufying the Philosophy of Freedom, to be guided to the 'place' where such an experience is possible for us, if we are willing, if we try.  Can we get as excited as Elon Musk about uncovering and meeting our own thinking nature, actually meet our thinking and find the source of our thinking?