Assignment #3 - The Chinese Room

John Searle's paper, "Minds, brains, and programs," attempts to show that:

1. A computer that is running a computer program is not understanding.
2. To be considered 'thinking,'  it will need to produce the causal features that the brain does.
3.  Strong AI will not succeed because they focus on programs, which can never produce thought.

To show this, he discusses a theoretical machine that can pass the Turing test. It can take in Chinese words and output a sensible Chinese response. He then says that this would be no different than a person, who does not know Chinese, sitting on the other end of the machine producing responses by following by hand the same program that the computer would follow. The user of the machine would see no difference between the machine answering them and the person answering them. This person does not know Chinese and would not understand what they were doing when responding, so this is not the same thing as thinking. If the person was not understanding when following the directions, then the computer would not be understanding either.

I might be over simplifying things, but I feel like this is very obvious. The way he showed it might be novel, but the idea behind it could have been proven through a survey. If you ask any person who has had a boring, mindless job (I am one of those people), they will tell you that they do not have to think or understand to perform the tasks. If the task are the same every time and are step-by-step, there is no need to understand. These same attributes apply to a program. Programs always handle a situation the same way and they are procedural. Therefore, implementing a program is not the same thing as understanding. Even without going into a hypothetical example like Searle did, I was able to come to the same conclusion.

He really made the paper more complicated than it needed to be. Of course just running a program is not the same as thinking. If so, all of our technology would be thinking on some level. There has to be something more to it than following a program, and if AI researchers want to make a truly thinking machine (and I don't believe most do), then they would have to figure out what that something is. I am not going to comment on whether or not I think it will be possible, but I truly have no idea.

The actual article probably could have been explained better and could have been condensed some.The Wikipedia article was much "better" (maybe not better in value but definitely better in explaining). It did have its moments when it would diverge into a short and related but not complete relevant side topic.

So to summarize, I didn't really care for how the article was written and I didn't think the conclusion was anything special. The way he came about the solution was novel but also more complicated than need be. I might just be oversimplifying it though. There were other things other than the Chinese speaking computer that he discussed, but to me, this was the focus of the paper (after all, the paper is even named after it). Hence that is what I focused on here.