Associations

     I haven’t got much, just now, but I hate to let this place get idle, especially after all the encouraging email and comments I’ve received lately, so I figure I’d drop at least a few lines.

     The arguments about artificial intelligence (AI) continue to rage. They may never end. A goodly portion concern the matter of consciousness or, alternately, sentience. These concepts are not definable in the Aristotelian fashion. For the moment, all we have are the assertions of living men that “I am conscious”… when they’re awake and responsive, at least.

     So we have humans, who have only a tenuous grasp on what it means to be conscious, arguing about whether software, driven by a rule-based architecture to assemble sentences from a combination of training and voluminous input, can ever be truly conscious. I can’t see an exit from this labyrinth. Frankly, we might as well debate whether a program can dream.

     All that having been said, I had an interesting conscious experience a little while ago. It may have some relevance to the debates over AI.

     I was going through my morning checklist of websites, which includes YouTube among others. That site presented me, no doubt “at random,” with a short video of a German musician, Carolina Eyck, playing and singing one of cinema’s most memorable bits of music: Ennio Morricone’s brilliant score to the three-way standoff that climaxes The Good, The Bad, And The Ugly:

     Extraordinary. I was once a theremin player, so I know how difficult it is to get exact notes out of that instrument. Add her voice to that, and it’s clear that Carolina Eyck has a great gift. But let me not digress.

     Eyck’s rendition of that memorable tune practically forced me to look for a clip of the scene itself. The cascade of memories that followed was too rich to describe in detail. Let it suffice to say that I relived much of the experience of long ago, when I first saw that movie. I felt blessed.

     Human memory is poorly understood. It’s unlikely to have any similarity to digital memory. In particular, surges of memory into one’s conscious thought stream are exceedingly difficult to explain. That may be because we lack the information required to trace the chain of associations that triggers them. Or it may be that they’re “random.”

     Randomicity is something else we have a hard time with. We observe certain phenomena in nature and, unable to place them in a firm cause-and-effect chain, we say they’re random. Are they? God alone knows. We do know this: we can’t generate “random” events. It’s not possible to do so consciously… and there’s that word again.

     Our efforts to know the world from top to bottom will always be limited by our own, innately human limits. Consciousness is one of the things that limits us. A conscious mind can never grasp a process that generates a random sequence of events, if indeed there is such a process. The two are antithetical to one another. In consequence, the word random will always be a label for things and events we fail to understand.

     But one thing we do know from top to bottom is how software operates… and AI is software. Software is inevitably procedural. Everything any program does can be traced to an originating input from which a chain of predictable decisions and actions leads to a predictable output. “Random” simply doesn’t apply to software. (No nonsense about “random-number generators;” you know perfectly well there’s no such thing.)

     That seems to exclude an AI from having the sort of experience I had this morning. The input, Carolina Eyck’s performance clip, was the fruit of a process that I choose, from my human limitations, to call random. The output was many-faceted, determined by my store of memories and my actions that followed. Even if we postulate that at some misty future time, an AI could be given a comparable experience, can we imagine that it would experience the sense of gratitude that came to me?

     I can’t see it – and that leads me to conclude that what we call “artificial intelligence” will forever lack the true gift of consciousness.

     Now, this may be mere human provinciality. Or it may be that our language, developed by conscious minds for the purposes of consciousness, is simply ill-fitted to grappling with phenomena that underlie consciousness. But if I’m seeing clearly, “artificial intelligences” will forever be distinct from human ones. They may be able to do things we can’t do, but I feel certain that the reverse is true and always will be.

     All that to one side, have an imagining: You, a great researcher and thinker, have discovered a remarkable phenomenon: two objects that produce streams of output that appear to be both completely random and completely identical. Can you think for a use for such a thing? Would using it require that you learn how to reproduce it? What if you never learn how to do that; could you make use of it then?

     Just a little early-morning food for thought.