Thursday, May 31, 2007

Can machines think?

It doesn't seem wrong to me to say that machines can think. The processing that happens to the visual or auditory signals within my brain before they reach my perception is very much like the same sort of thing that machines do. In fact, I could replace those parts of my brain with prosthetics and not notice the difference, I suspect. But it's unconscious, unperceived thinking. My brain is capable of doing a lot of unconscious (subconscious) processing of ideas, too. The difference is in a robot, all the thinking would be unconscious. You would have to use tricks, deceptions, to accomplish the things that are caused in humans by consciousness (perception and free will.) It isn't hard for me to picture that in the future, a robot could be designed to trick people into thinking it was conscious. But it would still be a trick. The Turing test wouldn't be useful in answering the question "Is a computer conscious?" but only "Can a machine think?" The answer to the second question will be "Yes" but that isn't what we really wanted to know.

Another argument: this web page contains thoughts, stored on a hard drive. If a system rearranges thoughts and produces new thoughts, that seems to me to be a good definition of thinking.


summerstay said...

When most people use the word 'think,' it encompasses consciousness of the thinking. If one wants to talk about thinking that is not conscious, one has to somehow point out that it is not conscious.

Darkmoon_UK said...

"The difference is in a robot, all the thinking would be unconscious."

On reflection, do you think this might be a gross assumption?