Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site ucbvax.BERKELEY.EDU
Path: utzoo!decvax!bellcore!ulysses!cbosgd!ucbvax!dewey.udel.edu!johnson
From: johnson@DEWEY.UDEL.EDU (johnson)
Newsgroups: mod.ai
Subject: Re: The Turing Test - A Third Quantisation?
Message-ID: <8603210932.AA02073@ucbvax.berkeley.edu>
Date: Fri, 21-Mar-86 04:33:08 EST
Article-I.D.: ucbvax.8603210932.AA02073
Posted: Fri Mar 21 04:33:08 1986
Date-Received: Sat, 22-Mar-86 20:14:18 EST
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 25
Approved: ailist@sri-ai.arpa


|Now,  supposing a system  has been built which  "passes" the test. Why
|not take  the process  one stage  further?  Why not  try to design  an
|intelligent system which can decide whether *it* is talking to machine
|or not?
|
|Gordon Joly
|ARPA: gcj%qmc-ori@ucl-cs.arpa
|UUCP: ...!ukc!qmc-cs!qmc-ori!gcj


Let me get this straight, a human cannot distinguish machine M1 from another
human, but machine M2 *can* distinguish M1 from a human. Will machines of type
M2 then debate about whether it is possible for a human to be modified to pass
the M2turing test?  Alternatively, perhaps M2s should try to create M3 s.t.
an M3 cannot be distinguished from a human by an M2, or how about an M4, which
is a machine that an M2 cannot distinguish from an M1?  But wait, how can an
M2 be sure that an M4 is not simply a copy of an M1?  Is some descendent of the
turing test a test that which tries to infer the nature of the designer from
the design?

-johnson@UDEL.EDU