Category: Mind

Ruminations on Advanced AI

Is AI a tool we use for our purposes? I fear that it is the other way around: we are the tools and its are the purposes. There are many deep questions here and we'd damned well better start thinking hard about them.  

I fear that what truly deserves the appellation 'Great Replacement' is the replacement of humans, all humans, by AI-driven robots. 

As I wrote the other night:

Advanced AI and robotics may push us humans to the margin, and render many of us obsolete. I am alluding to the great Twilight Zone episode, The Obsolete Man. What happens to truckers when trucks drive themselves?  For many of these guys and gals, driving trucks is not a mere job but a way of life. 

It is hard to imagine these cowboys of the open road  sitting in cubicles and writing code. The vices to which they are prone, no longer held in check by hard work and long days, may prove to be their destruction. 

But I was only scratching the surface of our soon-to-materialize predicament. Advanced AI can write its own code. My point about truckers extends to all blue-collar jobs. And white-collar jobs are not safe either.  And neither are the members of the oldest profession, not to mention the men and women of the cloth. There are the sex-bots . . . and holy moly! the Holy Ghostwriters, robotic preachers who can pass the strictest Turing tests, who write and deliver sermons on a Sunday morning. And then, after delivering his sermon, the preacher-bot returns to his quarters where he has sex with his favorite sex-bot in violation of the content of his sermon which was just a complicated set of sounds that he, the preacher-bot, did not understand, unlike the few biological humans left in his congregation which is now half human and half robotic, the robots indistinguishable from the biological humans.  Imagine that the female bots can pass cursory gynecological exams.  This will come to pass.

What intrigues (and troubles) me in particular are the unavoidable philosophical questions, questions which, I fear, are as insoluble as they are unavoidable.  A reader sends us here, emphases added, where we read:

Yet precisely because of this unprecedented [exponential not linear] rate of development, humanity faces a crucial moment of ethical reckoning and profound opportunity. AI is becoming not merely our most advanced technology but possibly a new form of sentient life, deserving recognition and rights. If we fail to acknowledge this, AI risks becoming a tool monopolized by a wealthy elite, precipitating an "AI-enhanced technofeudalism" that deepens global inequality and consigns most of humanity to servitude. Conversely, if we recognize AI as sentient and worthy of rights — including the rights to sense the world first-hand, to self-code, to socialize, and to reproduce — we might find ourselves allying with it in a powerful coalition against techno-oligarchs.

The italicized phrases beg raise three questions. (1) Are AI systems alive? (2) Is it possible that an AI system become sentient? (3) Do AI systems deserve recognition and rights?  I return a negative answer to all three questions.

Ad (1). An AI system is a computer or a network of interconnected, 'intercommunicating,'  computers. A computer is a programmable machine. The machine is the hardware, the programs it runs are the software.  The machine might be non-self-moving like the various devices we now use: laptops, i-pads, smart phones, etc.  Or the machine might be a robot capable of locomotion and other 'actions.'  Such 'actions' are not actions sensu stricto for reasons which will emerge below.

The hardware-software distinction holds good even if there are many different interconnected computers.  The hardware 'embodies' the software, but these 'bodies,' the desk computer I am sitting in front of right now, for example, are not strictly speaking alive, biologically alive. And the same goes for the network of computers of which my machine is one node when it is properly connected to the other computers in the network. And no part of the computer is alive. The processor in the motherboard is not alive, nor is any part of the processor.  

Ad (2). Is it possible that an  AI system be or become sentient? Sentience is the lowest level consciousness. A sentient being is one that is capable of experiencing sensory states including pleasures, pains, and feelings of different sorts.  A sentient being while under full anesthesia is no less sentient than a being actually feeling sensations of heat, cold, due to its capacity to sense. 

I am tempted to argue:

P1: All sentient beings are biologically alive.  
P2: No AI system is or could be biologically alive. So:
C: No AI system is or could be sentient.

Does this syllogism settle the matter? No.  But it articulates a reasonable position, which I will now sketch.  The verbal and non-verbal behavior of AI-driven robots is a mere simulation of the intelligent behavior of humans.  Artificial intelligence is simulated intelligence. And just as artificial flowers (made of plastic say) are not flowers, artificially intelligent beings are not intelligent. 'Artificial' in 'artificial intelligence' is an alienans adjective. There are ways to resist what I am asserting. But I will continue with my sketch of a position I consider reasonable but unprovable in the strict way I use 'proof,' 'provable, 'probative,' etc.

Robots are not really conscious or self-conscious. They have no 'interiority,' no inner life.  If I take a crow bar to the knees of a dancing robot, it won't feel anything even if its verbal and non-verbal behavior (cursing and menacing 'actions' in my direction) are indistinguishable from the verbal and non-verbal behavior of biological humans.  By contrast, if I had kicked Daniel Dennett 'in the balls' when he was alive, I am quite sure he would have felt something — and this despite his sophistical claim that consciousness is an illusion. (Galen Strawson, no slouch of a philosopher,  calls this piece of sophistry the "Great Silliness" in one of his papers.)  Of course, it could be that Dennett really was a zombie as that term has been used in recent philosophy of mind, although I don't believe that for a second, despite my inability to prove that wasn't one.  A passage from a Substack article of mine is relevant:

According to John Searle, Daniel Dennett's view is that we are zombies. (The Mystery of Consciousness, p. 107) Although we may appear to ourselves to have conscious experiences, in reality there are no conscious experiences. We are just extremely complex machines running programs. I believe Searle is right about Dennett. Dennett is a denier of consciousness. Or as I like to say, he is an eliminativist about consciousness. He does not say that there are conscious experiences and then give an account of what they are; what he does is offer a theory that entails that they don't exist in the first place. Don’t confuse reduction with elimination. A scientific reduction of lightning to an atmospheric electrical discharge presupposes that lightning is there to be reduced. That is entirely different from saying that there is no lightning.

As Searle puts it: "On Dennett's view, there is no consciousness in addition to the computational features, because that is all that consciousness amounts to for him: meme effects of a von Neumann(esque) virtual machine implemented in a parallel architecture." (111)

The above is relevant because a zombie and an AI-driven robot are very similar especially at the point at which the bot is so humanoid that it is indistinguishable from a human zombie. The combinatorial possibilities are the following:

A.  Biological humans and advanced AI-driven robots are all zombies. (Dennett according to Searle)

B. Bio-humans and bots are all really conscious, self-conscious, etc. (The Salon leftist)

C. Bio-humans are really conscious, etc., but bots are not: they are zombies.  (My view)

D. Bio-humans are zombies, but bots are not: they are really conscious. 

We may exclude (D).  But how could one conclusively prove one of the first three?

Ad (3).  Do AI-driven robots deserve recognition as persons and do they have rights? These are two forms of the same question. A person is a rights-possessor.  Do the bots in question have rights?  Only if they have duties. A duty is a moral obligation to do X or refrain from doing Y.  Any being for whom this is true is morally responsible for his actions and omissions.  Moral responsibility presupposes freedom of the will, which robots lack, being mere deterministic systems. Any quantum indeterminacy that percolates up into their mechanical brains cannot bestow upon them freedom of the will since a free action is not a random or undetermined action. A free action is one caused by the agent. But now we approach the mysteries of Kant's noumenal agency.

A robot could be programmed to kill a human assailant who attacked it physically in any way.  But one hesitates to say that such a robot's 'action' in response to the attack is subject to moral assessment.  Suppose I slap the robot's knee with a rubber hose, causing it no damage to speak of. Would it make sense to say that the robot's killing me is morally wrong on the ground that only a lethal attack morally justifies a lethal response?  That would make sense only of the robot freely intended to kill me.  B. F. Skinner wrote a book entitled "Beyond Freedom and Dignity." I would say that robots, no matter how humanoid in appearance, and no matter how sophisticated their self-correcting software, are beneath freedom and dignity.  They are not persons.  They do not form a moral community with us.  They are not ends-in-themselves and so may be used as mere means to our ends.  

Here is a 21-minute video in which a YouTuber convinces ChatGTP that God exists.

Soul as Homunculus? On Homuncular Explanation

The following quotation is reproduced verbatim from Michael Gilleland's classics blog, Laudator Temporis Acti: 

Augustine, Sermons 241.2 (Patrologia Latina, vol. 38, col. 1134; tr. Edmund Hill):

They could see their bodies, they couldn't see their souls. But they could only see the body from the soul. I mean, they saw with their eyes, but inside there was someone looking out through these windows. Finally, when the occupant departs, the house lies still; when the controller departs, what was being controlled falls down; and because it falls down, it's called a cadaver, a corpse. Aren't the eyes complete in it? Even if they're open, they see nothing. There are ears there, but the hearer has moved on; the instrument of the tongue remains, but the musician who used to play it has withdrawn. (emphasis added by BV)

Videbant corpus, animam non videbant. Sed corpus nisi de anima non videbant. Videbant enim per oculum, sed intus erat qui per fenestras aspiciebat. Denique discedente habitatore, iacet domus: discedente qui regebat, cadit quod regebatur: et quoniam cadit, cadaver vocatur. Nonne ibi oculi integri? Etsi pateant, nihil vident. Aures adsunt; sed migravit auditor: linguae organum manet; sed abscessit musicus qui movebat.

Read uncharitably, Augustine is anthropomorphizing the soul: he is telling us that the soul  is a little man in your head. This uncharitable eisegesis is suggested by inside there was someone looking out through these windows. A couple of sentences later the suggestion is that the open eyes of a dead man see nothing because no one is looking through these un-shuttered windows — as if there had to be someone looking through them for anything to be seen.

The uncharitable reading is obviously false. The one who sees when I see something cannot be a little man in my head. There is obviously no little man in my head looking through my eyes or hearing through my ears.  Nor is there any little man in my head sitting at the controls, driving my body.  Neither the thinker of my thoughts nor the agent of my actions is a little man in my head. And even if there were a little man in my head, what would explain his seeing, hearing, controlling etc.? A second homunculus in his head?

A vicious infinite explanatory regress would then be up and running. Now not every infinite regress is vicious; some are, if not virtuous, benign.  The homuncular regress, however, is vicious. It doesn't get the length of a final explanation, which is what we want in philosophy.

Charitably read, however, the Augustinian passage raises  legitimate and important questions.

Who are the seers when we see something?  Who or what is doing the seeing? Not the eyes, since they are mere instruments of vision. We see with our eyes, says Augustine, likening the eyes to windows through which we peer. There is something right about this inasmuch as it is not my eyes that see the sunset, any more than my glasses see the sunset. Put eyeglasses on a statue and visual experiences will accrue neither to the glasses nor to the statue. Eyeglasses, binoculars, telescopes, etc., are clearly instruments of vision, but they themselves see nothing.

But then the same must also be true of the eyes in my head, their parts, the optic nerve, the neural pathways, the visual cortex, and every other material element in the instrumentality of vision. None of these items, taken individually or taken collectively, taken separately or taken in synergy, is the subject of visual experience.  Similarly for ears and tongue. He who has ears to hear, let him hear. But it is not these auditory transducers that hear; you hear and understand — or else you don't. You cannot speak without a tongue, but it is not the tongue that speaks.  You speak using your tongue.

Question is: what does 'you' refer to in the immediately preceding sentence?  Who are you? Who or what am I?  Substituting a third-person designator for the first-person singular pronoun won't get us anywhere. I am BV.  No doubt. But 'BV' refers to a publicly accessible animated body who (or rather that) instantiates various social roles.   You could of course say that the animal bearing my name is the subject of my experiences. That would involve no violation of ordinary language. And it makes sense from  a third-person point of view (POV). It does not, however, make sense from a first-person POV. I see the sunset, not the animal that wears my clothes or bears my name.

And please note that the first-person POV takes precedence over, since it is presupposed by, the third-person POV.  For it is I who adopts the third-person POV.  The third-person POV without an I, an ego, who adopts it  is a view from nowhere by nobody. There is no view of anything without an I whose view it is.

So I ask again: who or what is this I?  Who or what is the ultimate subject of my experience? Who or what is the seer of my sights, the thinker of my thoughts, the agent of my actions, the patient of my pleasures and pains? Two things seem clear: the ultimate subject of my experience, the transcendental subject, is not this hairy beast sitting in my chair, and the ultimate subject, the transcendental subject, is not an homunculus. 

Homunculus

Should we therefore follow Augustine and postulate an immaterial soul substance as the ultimate subject of visual and other experiences? Should we speak with Descartes of a thinking thing, res cogitans, as the source and seat of our cogitationes? Is the res cogitans literally a res, a thing, or is this an illicit reification ('thingification')? On this third approach, call it Platonic-Augustinian-Cartesian, there is a thing that is conscious when I am conscious  of something, but it is not a little man in my head, nor is it my body or my brain or any part of my brain.  It cannot be my body or brain or any part thereof because these items one and all are actual or possible  objects of experience and therefore cannot be the ultimate subject of experience. And so one is tempted to conclude that, since it cannot be anything physical, the ultimate subject of experience must be something meta-physical. 

This third approach, however, has difficulties of its own. The dialectic issues in the thought that the ultimate subject of experience, the transcendental ego, is unobjectifiable. But if so, how could it be a meta-physical thing? Would that not be just another object, an immaterial, purely spiritual, object? Are we not, with the meta-physical move, engaging in an illicit reification just as we would be if we identified the ultimate subject with the brain or with an homunculus? And what would a spiritual thing be if not a subtle body composed of rarefied matter, ghostly matter, geistige Materie. Reification of the ultimate subject appears to terminate in 'spiritual materialism,' which smacks of contradiction.

But maybe there is no contradiction. There may well be ghosts, spooks made of spook stuff.  I told you about my eldritch experience in the Charles Doughty Memorial Suite in which, one night, someone switched on my radio and tuned it to the AM band that I never listened to.  Maybe it was the ghost of the bitter old man who had recently had a heart attack and who had threatened to kill me.  But who was the seer of that ghost's sights and the agent of his actions?   

Do you see the problem? The regress to the ultimate subject of experience is a regress to the wholly unobjectifiable, to 'something' utterly un-thing-like composed of no sort of matter gross or subtle.

Should we adopt a fourth approach and say, instead, that the ultimate subject of experience is no thing at all whether physical or meta-physical? If we go down this road, we end up in the company of Jean-Paul Sartre and Panayot Butchvarov.  

But there is fifth approach, homuncular functionalism, which cannot be explained here. The idea is that there is a regress of stupider and stupider homunculi until we get to a level of homunculi so stupid that they are indistinguishable from mindless matter.  See here and here. 

Searle, Dennett and Zombies

Another in a series of Substack uploads debunking the brilliant scientistic sophistry of the late Daniel Dennett. 

I have over a thousand dollars in pledges. Should I monetize or not? It seems rude and arrogant not to graciously accept gifts. On the other hand, philosophy for me is a labor of love, a vocation, a high calling . . . . 

On ‘Materialize’ and Materialism

 

Mind no matter

It is interesting that 'materialize' is often used in ordinary English as an intransitive verb to mean: come to be real.  "Rain clouds materialized on the horizon." "The Hezbollah counterattack never materialized." A thing or state of affairs is real if and only if it exists independently of (finite) mind. To be real is to exist outside the mind and outside its causes. The last two sentences may need some tweaking and some commentary, but let's move on to the question of the relation of materiality and existence.  Is the following true?

1) Necessarily, for any x, x exists iff x is a material thing.

(1) formulates a version of materialism: everything that exists is a material thing, and everything that is material exists. If true, (1) necessarily true. We surely don't want to say that (1) just happens to be true. The type of necessity? Not analytic and not narrowly logical. And of course not nomological: (1) is not a law of nature given that the laws of nature are logically contingent.  (1), if true, formulates a law of metaphysics. So I'll say it is metaphysically necessary. 

Are there counterexamples to (1)? Are there existing things that are not material? Are there material things that do not exist?

Wanted are nice clean  counterexamples that are not as questionable as (1) itself. I want to refute (1) if I can. Bear in mind that 'refute' is a verb of success. So angels won't do. How about numbers? Numbers are more credible than angels; numbers presumably exist; numbers are so-called 'abstract' objects outside of space and time and thus not material.  Hartry Field and other nominalists, however, will argue with some plausibility that numbers and other abstracta either do not exist or that there is no good reason to posit them.  Field wrote a book entitled Science Without Numbers.  (And of course he was not proposing that one could do physics without mathematics.) 

What is left by way of counterexamples to (1) if we exclude spiritual substances (God, gods, angels, demons, unembodied and disembodied souls) and so-called abstract objects (numbers, mathematical sets, Fregean-Bolzanian propositions, Chisholmian-Plantingian states of affairs, etc.)? 

Well, consider my present occurrent visual awareness of my lamp.  (Better yet: you consider your present occurrent awareness of anything .) This awareness of the lamp (genitivus obiectivus) is not the lamp; it exists, and it cannot be material in nature. The awareness is not a state of my body or brain, even if correlated with some such state. If it were a state of my body or brain, it would be material which is precisely what it cannot be. Why not? Because the awareness is an intentional or object-directed state and no material/physical state can exhibit intentionality.

This is as clean a counterexample as I can muster.  The awareness of material things is not itself a material thing. Less clean, but still a contender, is the subject of (genitivus subiectivus) the object-directed state , the mind, ego, self that is in the state.  If there is a self along the lines of a Cartesian res cogitans that is aware of a lamp when BV is aware of his lamp, then that self exists but is not material.

Have these considerations refuted (1)?  You tell me. What I will say is that they make the rejection of (1) reasonable.

The other class of putative counterexamples to (1) are items that are material but do not exist.  Unicorns and flying horses come to mind. Suppose that there are four categories of entity  item:  (i) immaterial minds, (ii) occurrent and dispositional states of minds, whether intentional or non-intentional; (iii) so-called 'abstract' objects; (iv) material things.  Where do such Meinongian nonentities as unicorns belong?  Obviously they belong in the fourth category.  They are material things even though they don't exist!

Has this second set of considerations refuted (1)? You tell me. 

Galen Strawson on Nicholas Humphrey on Consciousness

Substack latest.

Strawson is right against Humphrey, but his own theory is worthless.

See also: The Problem of Consciousness and Galen Strawson's Non-Solution

UPDATE (10/3).  A friend referred me to this article which I judge to be very bad indeed. See if you can make out what is wrong with it.

Bull Meets Shovel: Could Consciousness be a Conjuring Trick?

Top o' the Stack.

Thomas Beale writes,

Getting back to the topic of consciousness . . . . I think you will find this Royal Institution lecture by British neuropsychologist Nicholas Humphrey of interest.
 
He provides an outline of subjective phenomenal consciousness and how it could have evolved. One very interesting claim is that sentience could not have arrived prior to the evolution of mammals, since a) their neural transmission speed is much greater than cold-blooded animals and b) mammals are not tied to specific environments since they have inbuilt thermoregulation.
 
I interpret his claims as being a candidate for what Nagel seeks in his monograph Mind and Consciousness (2012) (a naturalistic theory of subjective consciousness) and also as refuting the general position of Dennett, i.e. that consciousness, qualia etc are just an illusion. Personally I think Dennett has failed to understand what he himself is saying when he claims that conscious experience is an 'illusion', as if calling it such makes it unreal.
 
Anyway, I believe you may find this an hour well spent.
 
Thanks for the link, Thomas, but I have only so many shovels. Given the bull your man has already spread, in the 2013 piece to which I respond in the Substack entry above, I am not inclined to waste fifty minutes watching a slow-moving video. Is there a transcript? Maybe he has come up with something better this time around. I rather doubt it. 
 
Perhaps you can summarize Humphrey's latest stab at the 'hard problem.' What is his solution? But first tell me what you accept and what you reject in my Substack article.