Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!think.com!mintaka!ai-lab!life!wlim
From: wlim@ai.mit.edu (William Y-P. Lim)
Newsgroups: comp.ai
Subject: A replacement for the Turing Test
Message-ID: <WLIM.91Jun25071848@rice-chex.ai.mit.edu>
Date: 25 Jun 91 11:18:48 GMT
Sender: news@ai.mit.edu
Distribution: comp
Organization: MIT Artificial Intelligence Lab
Lines: 20

The current discussion on the inadequacy of the Turing Test reminds me
of some ideas I had back in my PhD days.  Rather than go into a long
discussion on the topic (I've trying to get the full discussion
written down for quite a while; may be now it is time to finally get
it done), I propose that we use a very simple test:

    Have you and the "thing" in question play a non-violent,
    non-destructive game of "cat (you) and mouse (the 'thing')."  The
    harder it is to catch the "thing," the smarter it is.

Note: There apart from the requirements that game be non-violent and
non-destructive (i.e. we don't want anybody get hurt), anything goes
including lying, cheating, talking one's way out of the game, learning
and adapting to the other party, etc.


Willie

wlim@ai.mit.edu           (can read/post news)
wlim@gdstech.grumman.com  (can read but can't post news)