If we agree that it is *not impossible* to construct a conscious artificial intelligence, that is, one that *knows* and *feels* and *understands* stuff, how would we know when we got there?