Friday, May 28, 2010

Disputing John Searle's Chinese Room

I often hear people claiming Artificial Intelligence (AI) is not possible to create as proven by John Searle's Chinese Room argument. The details of that argument can be found here: http://en.wikipedia.org/wiki/Chinese_room. This argument has always seemed flawed to me but I could never quite put my finger on why until yesterday.

While reading yet another claim that AI is not possible as demonstrated by the Chinese Room it occurred to me that there is a fundamental assumption in the argument that is completely flawed. This assumption is that you can have an algorithm that appropriately responds to prompts given in a natural language without being intelligent. This is illogical. An algorithm is just a set of instructions to solve a problem. It's predetermined. It is not intelligent, it may exibit intelligence because it is the product of intelligence but it has no intelligence its self. I think this is John Searle's real point. If that's the case then we agree, algorithms are not intelligent even though they can exhibit intelligence.

Responding to prompts given in a natural language such as Chinese, English, etc. requires intelligence. This is due to the fact that intelligent beings create, maintain, and evolve natural languages for the purpose of communicating thoughts and ideas with each other. Words are added modified or discarded based on the common thoughts of a population of intelligent beings. Therefore something without intelligence will never be able to communicate correctly using a natural language because the language is always changing due to the intellectual evolution of a population. There is probably a much better proof for this so if this is the one point that people want to argue about I'm confident this can be proven to most peoples' satisfaction.

Now to my main point, the problem with the Chinese Room is that John Searle has used a contradiction as an assumption to base his arguments on. I think most people are familiar with the outrageous proofs that can be made when the proof starts with an assumption of 1=0, such as "I am the Pope". This is exactly what Mr. Searle has done. He has in effect said, "Take this thing that is not intelligent and assume it is doing something that requires intelligence." Once that assumption is made you can prove pretty much anything you want. The most obvious proof is, "There is no intelligence."