| It takes a lot of CPU power to simulate all the activites of a human mind, and most AI efforts have come a cropper simply because the computation was too complex. But some of the goals of early AI projects, such as speech recognition, are now commercially viable. Ultimately, we'll have to cope with Minsky's gedankenexperiment of some decades ago: how will you know, upon reading this, that it was constructed by a human as opposed to a computer, and if the constructor claims self-consciousness, can you object? |