
When I was young cognitive science caught my fancy. I guess my interest came from an earlier question..
what makes us human? What is the essence of human beings? And that in turn came from another question,
what is the human soul? Does such a thing exist? And to the question before that...
What, if anything, survives beyond death?. I guess it is
Descartes' Cogito Ergo Sum that made me decide what makes us human is that we can THINK.
What is the nature of thinking/ thought? If intelligence is the capacity to think,
what is intelligence? Do animals and
insects think the way we do? What if we could break thinking down to some rules and it could be mimicked? What if, God forbid, we could get
machines (not even living things) to think??? Would it be right to say
Organizations can be intelligent? If so, would they have a soul?
Years have gone by and my quest for the soul continues.....
So how do we know if a machine can think? Can following a few simple rules be genuinely called "thinking"? Over years several possible definitions have come up. Some of them require the machine to make mistakes (not be right all the time) while others require the machine to be able to solve new puzzles where there are no pre-existing instructions.. through learning. In 1951
Alan Turing suggested a simple test. If you can get a machine to respond in ways that mimic human responses to an extent that we cannot distinguish if its a human or a machine then it constitutes of thinking for all practical purposes. How many times have you met up with a chat bot on yahoo and flirted with it before you realized it was a machine after all? In 1964,
Joseph Wizenbaum said that this is a trivial program and wrote the program
Eliza to prove it. Eliza is a simple computer program that can hold seemingly meaningful conversations with you via a keyboard. Soon enough, despite the good professor's intentions, people began to pour out their heart to Eliza. She was programmed to be a therapist along the lines of
Carl Roger's Client Centred Counselling. (You can have a date with her
here.)
One of the tasks that require thinking is translation from one language to another. So here is a question. If a person is able to translate a text from English to Chinese, would that mean he understood the meaning of the text and was able to translate based on this understanding? John Searle (1980) in his famous description of the
chinese room does not think so. (Try out some chinese
here.)
All of this leaves us pretty much stranded without a clue what it means to say a machine or a human being can think. Sure we have a set of possible criteria, but none of them enough by itself. Perhaps there is no one criteria, but a mix, neither of them the answer but together, they capture a good part of what it is.
Wittgenstien called this language games.
But Artificial Intelligence or (AI) has made progress. Perhaps not in building robots that act and think like human beings, but in a number of small robots like the
roomba that cleans homes, or
AIBO the robotic dog. Robots have been used extensively in industrial production lines. Now there are microprocessors inside car engines, even refrigerators. And with the ability to talk to each other by WIFI or through wires, how much longer will it take for
skynet to become
self-aware?
If all this hasnt utterly bored you, you can always read
Scientific American's version of the Rise of the Machines.