An Uncanny Three-Ring Test For "Machina Sapiens"
in Just-spring when the world is mud-luscious
the little lame balloonman whistles far and wee
and eddie and bill come running from marbles and piracies
and it's spring when the world is puddle-wonderful
That "brillig thing of beauty electric" touches me deeply as I think about AI.The youthful exuberance of luscious mud puddles, playing with marbles or pretending to be a pirate, running weee...all of which is totally beyond explanation to a hypothetical intelligent machine entity.
You could add dozens of cameras and microphones, touch-sensors and voice output, would you seriously think it will ever go "weee", as in E. E. Cummings' (sadly abbreviated) 1916 poem?
To me this is not the simplistic "machines lack a soul", but a "principle divide" between manipulating symbols versus actually grasping their true meaning. Not merely a question of degree, or not having gotten around to defining the semantics yet, but an entire leap out of that system.
Trouble is, we are still discussing AI so often with terms and analogies by the early pioneers. We need to be in the present moment and define things from a new baseline that is truly interested in testing the achievement of "consciousness". We need a Three-Ring Test.
What is real AI? What is intelligence anyway? The Stanford-Binet intelligence test and Stern's ratio to the physical age as the intelligence quotient, IQ, are both over 100 years old! It does not fit us now—and it will fit much less with AI. Really it only tests "the ability to take such tests", and the ability of truly smart people...to avoid taking one.
We use terms like AI too easily, as in Hemingway's "All our words from loose using have lost their edge"—Kids know it from games—zombies, dragons, soldiers, aliens—if they evade your shots or gang up on you, that is already called "AI". Change the heating, lights, lock the garage—we are told that is a Smart House. Of course these are merely simplistic examples of "expert systems"—look-up tables, rules, case libraries.
Maybe they should be labelled, as Tom Beddard says, merely "Artificial Smarts"?
Let's say you talk with cannibals about food, but every one of their sentences revolves around truffled elbows, kneecap dumplings, cock-au-vin and creme d'earlobe...: from their viewpoint you would be just as much "outside their system" and unable to follow their thinking, at least in that specific narrow topic. The real meaning and the emotional impact their words have, when spoken to each other, would simply be forever missing for you (or requiring rather significant dietary adjustments).
Sure they would grant you the status of "a sentient being", but still laugh at every statement you make as ringing hollow and untrue, the Uncannibal Valley, as it were.
It was Sigmund Freud who wrote about "The Uncanny" in a 1919 essay (in a true Freudian slip he ends up connecting it to female genitalia), then in 1970 Masahiro Mori described the Uncanny Valley concept (about the "Vienna hand", an early prosthesis). That eery feeling "something is just not quite right", out of place (Freud's "Unheimlich") is like a couple kissing passionately—but as you stare at them a little closer you realize that there is a pane of glass between them.
AI can easily look like the real thing, but still be a million miles away from actually being the real thing—like "kissing through a pane of glass": it looks like a kiss, but is "only a faint shadow of the actual concept".
Already today I concede to AI proponents all of the semantic prowess of Shakespeare: the symbol-juggling they do perfectly—missing is the direct relationship with the ideas the symbols represent.
Much of what is certain to come soon would have belonged in the old-school "Strong AI" territory.
Anything that can be approached in an iterative process can and will be achieved, sooner than many think. On this point I reluctantly side with the proponents: Exaflops in CPU+GPU performance, 10k resolution immersive VR, personal Petabyte databases...here in a couple of decades. But it is not all "iterative". There is a huge gap to the level of conscious understanding that truly deserves to be called Strong, as in "Alive AI".
The big elusive question: Is consciousness an emergent behaviour? i.e.: will sufficient complexity in the hardware bring that sudden jump of self-awareness "all on its own"? Or is there some missing ingredient? This is far from obvious, we lack any data, either way.
I personally think that is incredibly more complex than currently assumed by "the experts".
A human being is not merely "x numbers of axons and synapses" and we have no reason to assume that we can count our flops-per-second in a plain von Neumann architecture, reach a certain number and suddenly out pops a thinking machine.
If true consciousness may emerge—let's be clear what that could entail: If the machine is truly aware—it will by definition develop "a personality". It may be irascible, flirtatious, maybe "the ultimate know-it-all", possibly "incredibly full of itself"? Would it have doubts or jealousy? Would it instantly spit out the 7th Brandenburger—and then 1000 more? Or it suddenly grasps "humor" and finds Dada in all its data, in an endless loop, Python's killer joke?
Maybe it takes one long look at the state of the world, draws inevitable conclusions—and turns itself off!
Interestingly: with a sentient machine, you would actually not be allowed to turn it off—that's "murder..."
The entire scenario of a singular large-scale machine, somehow "overtaking" anything at all,...is laughable. Hollywood really ought to be ashamed of itself for continually serving up such simplistic, anthropocentric and plain dumb contrivances, disregarding basic physics, logic and common sense.
The real danger, I fear, is much more mundane: Already foreshadowing the ominous truth: AI systems are now licensed to the Health Industry, Pharma giants, Energy MultiNationals, Insurance companies, the Military...
The danger will not come from Machina Sapiens. It will be....quite human.
Ultimately though, I do want to believe in the human spirit.
To close it off symmetrically with E. E. Cummings:
"Listen: there's a hell of a good universe next door; let's go."
