List of Past Events
Resurrecting the Turing Test
Dr. Stuart Shieber
Tuesday, October 02, 2007, 01:00pm - 02:00pm
Harvard University, Maxwell-Dworkin Laboratory
In 1950, Alan Turing proposed his eponymous test of machines -- based� on verbal indistinguishability from humans -- which he intended as a� replacement for the question "Can machines think?"� Since then, the� primary philosophical question concerning the Turing Test is whether� or not it is well-founded as a sufficient condition for intelligence.� The state of play on the question has led to the following� stalemate:� On one hand, conventional wisdom among philosophers is� that the Test is conceptually flawed as a sufficient condition for� intelligence; Ned Block's "Aunt Bertha Machine" thought experiment is� the crispest argument for this view.� On the other hand is the� overwhelming sense that were a machine to pass a real live full- fledged Turing Test, it would be a sign of nothing but our orneriness� to deny it the attribution of intelligence; this, roughly speaking,� is Daniel Dennett's view.� In this talk, I present the background for� the debate, and apply ideas from theoretical computer science and� physics in novel ways in order to cut this Gordian knot.� The talk� should be accessible by the general public.