Dartz Wrote:I can see them putting a 'subject' in it in a double-blind test, throwing up a negative result.... only to find that it was in fact an ordinary person off the street. Back to the drawing board.The Turing Test is arguable a test of role-playing (or, more crudely, lying).
Ellison Voight and Allen Kampff have a lot of work left to do.
http://en.wikipedia.org/wiki/Turing_test
Can your AI role-playing being a human?
That sounds better than "we're designing AIs to lie so well humans can't spot them doing it".
But, your point is well taken. Something might be intelligent, but do they have the "social intelligence" to get on with a human society? Part of this is empathy and the ability to (correctly) model the behaviour of other intelligences. Then, there is the matter of being concerned enough about those other intelligences to cooperate with them in cases when you have spare resources left over from what you need to survive.
Maybe that is the "Give A D*mn" Test? [grin]
(That is the religious version of the test, by the way. [grin] )
--
"It is the business of the future to be dangerous" - Hawkwind