IT Curmudgeon

Alan Turing was a brilliant mathematician and philosopher who considered the ramifications of true thinking machines. In the 1950's and 1960's, people were discussing whether machines would ever be able to think.

In one of his theoretical proposals Alan Turing devised what is now known as the Touring Test in which he poses the question, "Can a machine make us believe it is human?"

The test postulates a scenario in which a person is conversing with two "entities," and if the person cannot tell which entity is human and which is a machine, then the machine will have passed the test.

This test is incredibly challenging when you consider a question like, "How about those Sox?" or "How can we stop the Samali pirates?" or "What is a three letter word used by Homer Simpson?." This requires contextual information to understand the question and to postulate a response.

In 1997 IBM created a computer program called "Deep Blue" that defeated world champion Garry Kasparov. That was impressive because the human brain is very efficient at finding non-linear solutions to problems.

Now IBM wants to develop a program called "Watson" that will compete in the television game show Jeopardy. If you are familiar with the show, answers often require contextual world knowledge and sometimes a sense of humor. Even if you can't give a solid definition of "artificial intelligence" or "AI" you can tell that this is getting pretty close.

"Artificial intelligence" is a term that has been both misused and over-rated. One misuse of the term is when it is applied to an "expert system." An expert system is one in which rules are programmed to assist humans apply complex rules efficiently. Spelling and grammar assistants would be examples of expert systems, especially if they work properly ("How about those socks?").

People offer different definitions of artificial intelligence, but it often boils down to the idea that a machine can solve problems for which it has not been programmed. That is, it can find new solutions to problems. As time passes, we are getting closer to being able to both refine the questions and develop these kinds of systems.

As we find our way to making computers smarter, we are finding out more about what it is to think and be intelligent, and what it is to be human. In this way we are learning more about ourselves. James Joyce, in the book "Ulysses" said that looking for consciousness is like "looking for darkness with a flashlight." Hopefully projects like Watson will reveal more about us and we can move the self-knowledge meter a little farther forward.