Reading #??, The Chinese Room Question

Comments:
http://chiblog.sjmorrow.com/2011/01/reading-2-chinese-room.html
http://shennessy11.blogspot.com/2011/01/chinese-room.html

I presume a reference section is not required for this post.

The Chinese Room Question is a philosophical assertion with respect to the AI field. The assertion is that an individual who knows no Chinese characters sitting inside a closed room (a set up similar to the Turing test, which I presume the reader to be familiar with) manually executing a computer program using cards to reply to inputs that are likewise Chinese characters printed on cards. The individual is then able to converse in Chinese without actually understanding the language. The argument is that this proves the impossibility of "strong AI"; that is, the argument says that this scenario demonstrates that a computer can never be built into a true mind, but only a simulation of one.

While I actually tend to concur with the conclusion, I disagree that the Chinese Room demonstrates it. The Chinese room does not prove that the individual does not understand Chinese any more than it would prove that he does. From the perspective of an outside observer feeding in cards, the output would of course still be correct if the operator did know Chinese.

What the Chinese room does succeed in doing is demonstrating the Turing Test is not adequate to demonstrate strong AI. This is accomplished by showing that that thought experiment, viewed in the perspective of a test for intelligence, can be fooled by an a "program" that does not actually understand the inputs but rather only simulates it.

From the website of Glasgow University's Philosophy department, via GIS