Has a machine really passed the Turing test?


When I first took a Turing test.

I remember vividly when I was introduced to the Turing Test in the early '70s. I was walking around Bear Lake in Estes Park, Colorado, with a senior, distinguished colleague in the field of vision research, an affable and delightful man named Fergus Campbell. He was in Boulder for a conference and I was excited that he would spend time with me, a brand-new Assistant Professor.

I don't recall how we got onto it, but he started to ask me questions as if I was taking part in a Turing Test. I didn't catch on and went right along "down the garden path" of his gentle questioning. Finally, I said something like, "Wait a minute, you mean a computer is just like a brain?" He responded by asking if that wasn't just what I had proved for myself in answering all his logical questions.

There's more to it, of course.

Real Turing Tests are tried at research centers all the time. In fact a popular example of what might be called a limited Turing Test was played out in 1997 when the computer known as "Deep Blue" (from IBM) defeated the reigning world champion chess Grand-Master Karpov in a seven game match.

So, has any computer successfully passed the full Turing Test in a rigorous, real-life situation? No, although there are still those who would answer, "Not yet." Of course, in the difference between these two answers lies the topic of this essay, "Is the brain (like) a computer?"

The 'Not yet' camp suggests the problem is one of practical implimentation and relies on better programming techniques and bigger and faster computers to solve the problem. Those who believe that "No" will always be the answer raise theoretical issues that will be taken up on other pages of this essay.

However, as with any theory based in mathematical formalism, such as Turing's, there are arguments over the applicability of the details to human behavior, such as the fact that the machine may take an infinite time to complete a particular task.

Just how is consciousness judged?

If you could teach a monkey to play a creditable game of checkers and it displayed strategic thinkiing would you attribute consciousness to that monkey? Well, there are computers and their program that can always win at checkers . . .