“In order to make an apple pie from
scratch, you must first create the universe"
- Carl Sagan
My computer is what is known as a tough room.
I mean that in the stand-up sense that it never seems to get my jokes. It transcribes them well enough, stores them, transmits them, even translates them…but nary a guffaw. It is great at handling gigs but not so much giggles.
I know how dumb that insight sounds, but humor is at the top of every Gizmotician’s list of must-have traits simply because it is so human. Like the researchers at the University of Cincinnati who built a bot that gets puns. It is really a language parser that finds ill-fitting words and when a similar sounding word fits better into the sentence, it decides it has found a joke. Or the folks at the University of North Texas who created a program that flags sentences that have any jokish word in them…like the word “drunk.”
But all that is pretty petty stuff. A computer that really knew funny would be a breakthrough not just in machine comedy but in artificial life, because humor may be the last bastion of human pre-eminence now that chess and calculus have been conquered.
To accomplish that you would need something beyond an algorithm (and surely beyond Al Gore Rhythm for that matter…bah-dum.) to actually groan at a pun or hoot at a jab. To wit: wit. In fact you would need some kind of A-Con – Artificial Consciousness – because humor involves our understanding of ourselves as beings in the world and our place in it and therefore why things do not work out. After all, that is what most humor is based on. Aggravation, irony, outright rage…these are the wellsprings of the hahaha and I know of no lab in the world working on that.
One reason for this might be the notion that consciousness – on any level – is impossible to recreate in the artificial realm of the Gizmos. The two are incompatable. In this view, it is not enough for computers to do what they do so well…compute really really fast. Consciousness is not computation, no matter how swift, and never can be.
Take the problem of the Chinese Room for example.
This is a thought experiment like Castel’s Ocular Harmonium or Turing’s Machine. The original version of the Chinese Room problem involved imagining a man in a room but let us consider a more updated version. Suppose you had a chip implanted near the auditory center of the brain that could instantly take any phrase said to you in Chinese and forward the signal to your cortex in English. A translation chip, in other words, assuming of course that you only spoke English. Pretty neat, but could you then say that you understood Chinese? The chip is simply a cool device, making computations not sense. True understanding and knowledge rely on consciousness; you need self-awareness to have true knowledge of.
Or do you?
If you can hear Chinese and understand it in your own tongue, so to speak, isn’t that knowing Chinese? The computation people don’t think so. Knowledge is something deeper than data, consciousness more profound than merely manipulating it.
The opposing view is not so strict.
It suggests that consciousness is not a single thing, on or off, have it or don’t. Instead it can be seen as a vast scale of possibility ranging from a slug’s mere glimmer of sentience all the way up to Emmanuel Kant. Those folks see knowledge the same way too. Computers “know” in a limited way; we “know” in a more expansive way. In this sense, you with your chip do “know” Chinese in some sense and on some level because knowledge is variable, just as consciousness is.
So maybe those rudimentary punbots and drunkbots are a step in the right direction towards something like Tummler 5.1 – a low brow comic entertainment AI – that would bother asking if the name Pavlov rings a bell…bah dumbum. Or even better would not only know that “eleven plus two = twelve plus one” is correct but smile when it realized that the two statements are also perfect letter anagrams of each other.
Or know to pause in the middle of “take my wife…please” not because there is an ellipsis there but because that is where the fun is.
Humor, cleverness, the appeal of hardyharhar…these are all moves towards a conscious gizmo, which is clearly were all this is heading. And it is a worthwhile endeavor because, as Hermann Hesse said, eternity is a mere moment, just long enough for a joke. After all, when you think about it, on one level the only difference between even the lamest comedian and Emmanuel Kant is mostly that Kant was not all that funny.
But then again, he didn’t have an open mike night.
Ba-dum.
No comments:
Post a Comment