On Fri, 10 Sep 2004 09:10:19 +0200, Vincent Archer said: > Emulating a human is very very different from making a sentience. That's > the main flaw of the Turing's test: it attempts to prove the existence > of human-type sentience, not sentience in general.
Douglas Adams understood this one - "So long, and thanks for all the fish..." We usually restrict ourselves to human-type sentience because we have a very low number of examples of alternatives (outside of the occasional science fiction writing that posits the existence of a hive mind or a sentient, thinking boulder - note that such are invariably a human's thoughts on what said thing *might* be like, and as such may not reflect a real hive mind or boulder). The basic concept of the Turing Test works for alternate sentiences as well - there's no real reason why it couldn't be (for instance) a human trying to tell the difference between a hive mind and a computer pretending to be a hive mind, or a hive mind trying to distinguish between a hive mind and a computer. The difference in the two is only in that a hive mind is more likely to be able to think of "trick questions" than a human is, so a computer pretending to be a hive mind is more likely to fool a human than another hive mind. But we already know this innately - you can pretend to be a Hasidic Jew, and probably fool somebody in Montana who's never seen or heard of a Hasidic Jew before. But trying to pull the the same stunt on the front steps of a synagogue in Israel....
pgprou8F60Ey8.pgp
Description: PGP signature