The Chinese Room experiment is an attempt by philosopher John Searle to prove that machines can’t be smart. It involves a person in a room pretending to be a computer. Here it is: You don’t know Chinese. You’re put into a room. You are given voluminous instructions. Every so often, somebody slips into the room a sheet of Chinese writing. You carefully follow the instructions, not knowing what you’re doing, making long calculations, and finally end up with another sheet of Chinese writing which you slip out of the room. From the point of view of people outside the room, they’ve asked a question in Chinese and gotten a sensible reply, so they might figure that this room is intelligent. But if that’s so, says Searle, then where’s the intelligence? It’s not in you; you’re just following instructions and you don’t know Chinese. And it’s not in the instructions; they didn’t do anything. So Searle concludes that it isn’t anywhere and the room isn’t intelligent, it’s only faking it and that doesn’t count.
I personally think that this argument is stupid, so I may not have put it the strongest way. It’s like asking: Is the behavior in the computer’s processor or in the computer’s program? Neither, you need both. To look at it another way: A simulated car is not a car; it’s a simulation. But simulated addition is addition; addition is something you do with information, so if you simulate the steps of doing addition then you are doing addition. Being smart is also something you do with information, therefore simulated intelligence is intelligence. You can’t fake it.
updated 28 June 2000