Tuesday, 20 December 2016

Robots, AI, and what consciousness is.

I've been watching HUMᗄNS, a BBC show, and also Westworld, on iTunes. Both are about AI androids who start exhibiting human characteristics. But the central premise on which these shows are based, that a machine that appears to be human actually is, is not really explained - it is simply assumed. The central premise is that same miracle in the Golem myth - that somehow a robot/machine/lump of clay can be given human/divine breath and made to live.

It started me thinking - the central characteristic of a human being is self-consciousness - I realised a long time ago that Descartes' fundamental basis for all of his philosophy of a Clockmaker universe, "I think therefore I am", is actually illusory - "I am" is the more fundamental realisation. That this realisation is possible for us is explained in Talmudic and Biblical teaching by God having breathed His life, a portion of His spirit (breath and spirit being the same word in Hebrew, neshamah) into us. In ontological terms, it follows, "I am, therefore I think". (i.e. The fact that I am is what causes me to be able to think)

A computer in its own way 'thinks' - the circuits bleep over and create a certain pattern - but the meaning of that pattern only becomes apparent to a human being. In fact, the meaning of any pattern requires a conscious being to see it, for it to have meaning. In other words, the genetic code is only the mechanics of a machine that orders the process that creates the body, but it must have been designed, because it is a construct involving symbolic meaning. i.e. it is a code - but strangely enough, the only living creatures who can see that it is a code are human beings (or other sentient beings, angels, God, aliens perhaps). The bacteria who benefits from this vastly clever piece of engineering cannot see that the DNA through which it comes into existence is a symbolic representation of the creature that comes into existence. (This awareness of symbolic pattern, by the way, is another aspect of self-awareness - the ability to discern the difference between signifier and signified - something post-modernism minimises by destroying that difference, being a system that attempts to treat human consciousness as analogous to a computer in an endless goto loop - which is precisely what consciousness is not. The Incompleteness Theorem of Kurt Gödel demonstrates that ultimately it takes a human being to discern when a theorem is self-referential, i.e., in an endless goto loop. Similarly, despite programmers putting in super-programs called daemons to monitor the program while it is running, and daemons to monitor the daemons, they still can't stop the computer from crashing from time to time, getting itself into an endless loop. You see, a machine is only a machine, but the Spirit of God is the Spirit of Truth. )

There are already dolls that talk to children, mobile phone 'personalities' (Siri), games that can hold conversations with us, and these can be programmed to have apparent feelings. I am sure Google eventually will become very good at extrapolating normal human emotions from the massive amounts of data they have collected about people - perhaps to the degree that they can eventually create an AI machine that can realistic imitate a human, and fool many people into thinking it is alive and conscious.

But if there is no centre to that creation, no part of it that can honestly say, "I am", then despite its apparent 'feeling', 'thought', even despite the fact that it might appear to have self-awareness, if there is no real self-consciousness, no divine spirit, then it is still nothing more than an artificial machine imitating human behaviour, something electronic, certainly, that really could just as well be made of cogs and wheels. We don't have to worry about offending it, hurting it, breaking it, sinning against it, except to the degree that we might be hurting its owners or ourselves by doing such things, just the same as any other machine. (Stealing a car is wrong, not because the car has feelings and sentience, but because the owner does. )

It seems to me that self-consciousness is spiritual, an effect of having a spirit. It is an exceptional thing for an animal to be able to recognise itself in a mirror. It is a normal thing for a human to be able to do so. Having feelings also, i.e. a soul, (nefesh) is a miracle in itself - only certain animals have this almost to a human degree - the higher mammals. But self-consciousness is uniquely human.

Some people, who insist that we are nothing more than machines ourselves, would say that the existence of self-awareness is an illusion. But the fact is, the grammar of their own statement shows how ludicrous this statement is: for there must be someone having an illusion, for it to be an illusion. If there is no self, if we are machines just as much as a computer is, then there can be no illusion at all, because there is no real person at the centre of it all who can be deceived.

I think self-consciousness in AIs is a problem that is insoluble to science. I really don't believe a human being can create another self-conscious being except through the usual methods (a sperm and an egg).

(In Henry Lawson Hero of the Robot Revolution, (*spoiler alert this paragraph*) of course, I solved the problem of the soul and spirit in AI machines in the narrative by giving them a soul, a spirit, that has been transferred from a human by some unexplained and drastically inhuman process. This is the only way I believe that AIs could be truly conscious, if a spirit could be transferred somehow from one vessel to the other. This in itself is a scary thought and is the main source of the horror in my novel, for anyone who has read it.)

In these shows (HUMᗄNS, Westworld, etc) the terrible thing that happens is that these AIs whom anyone with any human feeling can see are conscious, are being treated as impersonal machines, and they are fighting back. If this was possible it would indeed be terrible. It is the premise of many, many science fiction stories, from Frankenstein to Rossum's Universal Robots, to Blade Runner.

But no one yet has shown that it is possible to create a conscious machine, and no one yet has even shown us what consciousness is. A machine may be able to be created that can feign consciousness, but until this mystery can be solved (which I believe it cannot, precisely because self-consciousness is one aspect of the divine image in us, perhaps the main aspect, and this is a mystery) - until this mystery can be solved, we will never have to worry about offending, hurting, sinning against the machines in our lives. No, we should worry far more about hurting our fellow humans, and the other living creatures in our lives - a dog, after all, while not self-aware, does have genuine feelings (Nefesh) but a computer never will.



No comments:

Post a Comment

Dear griffin friend and dirigible traveller - please feel free to leave a comment. I always try to moderate comments within 24 hours or earlier, so hopefully it will appear soon.