John Searle addresses the point of the ability of Artificial Intelligence (AI) to understand, in Mind Brains, and Programs. His main argument is that because AI’s are computers and computers have no thoughts of their own, they cannot understand. Any actions being performed to simulate behavior are confined by the programs available to the computer. He presents the example of a man linking Chinese characters and appearing to know the language, but in reality the man is just following the instructions given to him ( the program).
This example serves well to explain how although a computer can look like it understands a story, it can do no more than “go through the motions. ” Of course such a definitive standpoint on an issue as controversial as the capacity of an AI to understand will draw many critics. The criticism of his theory that I find to be the most credible is The Other Mind Reply offered by Yale University. This line of thinking asks: if behavior is what we can determine the presence of cognition through, and an AI passes a behavioral test, why don’t we attribute cognition to it?
I myself do not believe in the philosophy of AI understanding, because to support either side on this issue one must have a belief for or against the ability of man to create another being capable of thought. I do not believe that any machine based creature we may ever create has the ability to think. Thought is something that is independently done and cannot be given to another, or more accurately, programmed in. Regardless of however many tests that may be passed simulating thought and understanding, a programmed being is not capable of thought and understanding.