John Searle, answered from the texts and cited to the page.
Strong artificial intelligence — the thesis that the appropriately programmed computer literally has cognitive states and that the programs explain human cognition — is false. The Chinese Room thought experiment shows that a person manipulating Chinese symbols by formal rules without understanding Chinese instantiates exactly what the programmed computer instantiates, and produces no understanding. Syntax is not the same as, nor is it by itself sufficient for, semantics.12345
The point is logical, not empirical: no amount of formal-symbol manipulation, however complex, can constitute the having of mental content. The Systems Reply (the room as a whole understands), the Robot Reply (give the system sensors), the Brain Simulator Reply (simulate neuron-firing), and the Combination Reply each fail for the same reason — none of them adds semantics to the syntax, they merely add more syntax.
It does not matter how well the system can imitate the behavior of someone who really does understand, nor how complex the symbol manipulations are; you cannot milk semantics out of syntactical processes alone. In order to refute the argument you would have to show that one of those premises is false, and that is not a likely prospect.The Mystery of Consciousness, pp. 13–14
This is why I can, at least in principle, carry out the steps in the program in the Chinese Room, even though my 'hardware' is quite unlike that of commercial computers.Philosophy in a New Century, p. 68
There you are locked in your room shuffling your Chinese symbols and passing out Chinese symbols in response to incoming Chinese symbols. On the basis of the situation as I have described it, there is no way you could learn any Chinese simply by manipulating these formal symbols.Minds, Brains and Science, pp. 33–34
You understand the questions in English because they are expressed in symbols whose meanings are known to you. Similarly, when you give the answers in English you are producing symbols which are 33 "a —a — meaningful to you. But in the case of the Chinese, you ha' none of that.Minds, Brains and Science, p. 34
Hence there are two things. The urge of the reductionist to try to hang on to the reduced notion amounts to a quiet abandonment of the reduction. If he really were consistent in the reduction of mental states to program states, he should have said, "There isn't any such thing as understanding in addition to symbol manipulation, there is just the symbol manipulation.Philosophy in a New Century, p. 70
Held without modification across thirty-five years and multiple book-length defenses. The argument distinguishes strong AI (computers literally have minds in virtue of programs) from weak AI (computers are useful tools for modeling cognition); Searle has no quarrel with the latter. The position is one of the most-attacked and most-defended in recent philosophy. The argument is logical, not a wager about future technology — no new development in machine learning can make a syntactic engine semantic, because the difference is conceptual, not engineering.