Depth Perception / Time Perception
He’s tough slogging and not nearly as famous or as well read as he once was, but Alfred Korzybski remains one of the most innovative and insightful thinkers of our age.
There’s a lot to be said about General Semantics, but one core idea is the concept of a time sense, and how a time sense marks a difference between human and non-human intelligences.
Basically, Korzybski wrote that human beings were unique insofar as they could enter into contracts based on a sense of time, of something happening in the future. “If you bring me two bushels of corn next harvest, I’ll give you one of my goats.”
Notice this is not a simple linear bargain (“I will give you a goat for two bushels of corn right now”) but a promise to do something if an even occurs in the future.
Animals, Korzybski reasoned, could not enter into contracts because they could not imagine a future. To them, everything happens in the immediate “now”: There either is / is not corn that could prompt a trade.
(Truth be told, since Korzybski first came up with his idea, there has been ample evidence to indicate animals are more than capable of planning ahead. Nonetheless, his original proposal stands: Animals are animals and not humans because they do not experience the passage of time.)
This could be a phenomenon akin to binocular depth perception.
We “see” depth if we have two eyes and a functioning occipital lobe.
If we lack either vision in one eye or the ability to process binocular images in our heads, we would see the world as a “flat” experience.
(We could still use a wide variety of visual clues to judge distance but we would not see depth.)
Depth, that impossible to describe sensation of visualizing the space between two or more objects (as opposed to merely intellectually understanding there is space between them) is a real phenomenon that can only be emulated, not duplicated in mechanical form.
To a similar degree, we experience the passage of time as a continuum of past-present-future, but to animals and to AI that experience is utterly meaningless.
They grasp certain clues that help them arrange things chronologically but they have no intrinsic sense of same.
They -- animals and machines -- experience time only as “now”, Korzybski would argue, though they can emulate “then”.
We know there are people like this as well: Folks who are incapable of planning ahead or executing a timetable because they simply cannot process a concept of “the future” and only respond to immediate stimuli.
(How much is merely poor impulse control / failure to accurately gauge consequences, and how much is a genuine deficiency of experiencing time is a task for individual diagnosis.)
AI behaves the same way: It can only respond to an actual stimulus in the here and now, it can’t really look to the future to plan things so it cannot, in a very real sense of the term, make a mistake.
AI’s response to any situation is purely binary, there can be no weighing of stimuli to determine the best course of action.
“Planning” by AI consists of a set of programmed responses to anticipate stimuli: When A, do B; when C, do D, etc., etc., and of course, etc.
To lack a time sense is to not be fully human.
© Buzz Dixon