🚀 go-pugleaf

Newsgroups: comp.ai
Subject: Re: Turing Test and Subject Bias
Summary:
Expires:
References: <>
Sender:
Reply-To: oded@wisdom.bitnet (Oded Maler)
Followup-To:
Distribution:
Organization: Weizmann Inst. of Science, Dept. Of Math, Rehovot, Israel
Keywords:


In article <> gilbert@cs.glasgow.ac.uk (Gilbert Cockton) writes:
>
>Then how valid is the Turing Test?
>
>Just what sort of Science did young Mr. Turing have in mind when he
>decided that subjective opinion could ever be a measure of system
>performance?
>
>How do AI types *REALLY* test their systems?
>--

The essence of Turing's test is in the notion of INDISTINGUISHABILITY
by an observer. There is a machine in Glasgow that posts
from time to time some sequences of characters that succeed in fooling
me (even when I'm sober) so that I cannot distinguish it from a human
local-patriotic social-science-type :-{}. Such a machine passes the Turing
test. Maybe a stronger observer than myself (exponential, infinite) would
have caught this imposter.

Of course this is not "science" unless I augment it with statistical
"objective" gadgets.

BTW, Turing's test ideas had a lot of influence (so my crypto friends tell
me) on various notions in Cryptography and Complexity such as interactive
proofs and zero-knowledge protocols. But there you have to distinguish
between someone who knows some fact and someone who doesn't, or between
random and non-random, and not between intelligent and non-intelligent.


        Oded Maler
        Department of Applied Mathematics
        Weizmann Institute of Science
        Rehovot 76100, Israel
        (oded@wisdom.bitnet)