Will any android ever achieve the grace and fluidity of the human organism--not just in physicality but in the electro-chemo-mechanical functions known as thought and emotion? Certainly they can approximate or mimic the human organism--but can they ever achieve the almost water-like nature of a human? Hmm, maybe that is the key to perfect AI: liquid crystality.
I once perused a book called The Society of Mind by cognitive scientist Marvin Minsky. There he talked about how all of the simple functions--push and pull, move and stop, and so on--work in concert to behave in unpredictable and creative ways. Perhaps the key to AI is determining the simplest functional substances that can work together.
A corallary to the "society" concept of the individual mind is that of the group. A giant anthill was once pumped full of cement, allowed to dry, and then excavated. The scientists discovered a complex "metropolis" of highways and biways, ventilation and waste management systems, incubation rooms, and so on--all seemingly designed by a single architect. No one ant knew what was going on on the macrocosm, but by God the thing worked as a whole, and each ant did his mighty little part.
That is how human civilization works, as well. Of course. What else but compartmentation and an invisible "hive mind" can account for the fact that we all know as individuals what the collective should do, but the collective doesn't seem to give a damn what our puny individual thoughts are. No matter how influential an individual, the whole will move in ways much more similar to a flock of birds or a weather system: unpredictable, fluid.
There it is again: fluidity. So an organism of artificial intelligence must be fluid, no? How can we create an artificial human that thinks, feels, and moves with such grace as that normally afforded to humans--or to animals--or to the way a forest grows up and self-regulates and achieves homeostasis?
Perhaps it is time to re-engineer the very concept of intelligence. Why not a stone have intelligence? Why not a mote of dust? Our galaxy is in itself seemingly "of some design"--the way it spirals, pirouettes--like a ballerina, or a dolphin at Sea World. Why not every level of existence be afforded some "intelligence"? Not only will this go a long way towards improving our understanding and respect and empathy for the world around us, but it could help to redefine the quest for a perfect AI.
Maybe we should redefine what it means to be intelligent before we try to create a robotic being that, for all intents and purposes, is a human being in a very real way. Instead of trying to create an artificial homo sapiens sapiens, why not create an artificially intelligent substance or fluid? Maybe it has already been done. Maybe we are surrounded by artificial intelligence already. Orville and Wilbur Wright created an airplane that captures air currents and manipulates air pressures in such a way as to create lift. The human arms cannot do that--and yet the seemingly unconscious airplane wing is impressive and awe-inspiring just the same.
Back to ants: I saw an article in Popular Science recently about real-world "zombies"--or beings whose brains have been compromised by some outside force. There is a fungus that attaches itself to an ant's cranium, injects it with chemical "commands", rides the ant to an ideal location for the fungus to reproduce, and then boom: the fungus bursts out of the poor ant's skull having been incubated therein, in just the right place for its spores to take flight and find more ants to use for procreation.
Pretty gruesome stuff, but besides that, what can we learn from that? What can we learn about the cooperative nature of "intelligence"? Certainly a fungus is not intelligent like a human. But that does not mean it is "lower" than a human (unless we are talking about actual spatial relationships, in which case, sure, yes, it is physically lower towards the ground.) Instead, the fungus is intelligent like a fungus.
We must meet intelligence on its own terms, not on the arbitrary narcissistic terms we set out for ourselves. Yes, it is natural to see the world in terms of what we can most immediately relate to--out own human incarnations, in this case--but why not expand the very definition of the self? Why not say, "Yes, that stone is me, that ant is me, that fungus is me, that galaxy is me." Of course, we don't have to mean this in every literal sense, but perhaps if we simply imagine it to be true for a moment, we can empathize with beings that can help us to increase our understanding of Things.
Fluidity. Self. Cooperation. Unconscious organization giving rise to consciousness. Emergent properties. All of these concepts and more must be applied to artificial intelligence--not just mechanics, electrodes, and software. Why limit ourselves? Perhaps the great breakthrough in artificial intelligence will come not from a tinkerer, not from a scientist, but from a jack of all trades--or a society of jacks of all trades.
These are some of my unpremeditated thoughts on artificial intelligence.