Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: Notesfiles $Revision: 1.7.0.10 $; site uiucuxc
Path: utzoo!watmath!clyde!cbosgd!ihnp4!inuxc!pur-ee!uiucdcs!uiucuxc!bantz
From: bantz@uiucuxc.CSO.UIUC.EDU
Newsgroups: net.ai
Subject: Re: Two AI software packages
Message-ID: <28900016@uiucuxc>
Date: Sun, 9-Feb-86 10:45:00 EST
Article-I.D.: uiucuxc.28900016
Posted: Sun Feb  9 10:45:00 1986
Date-Received: Tue, 11-Feb-86 07:13:51 EST
References: <842@aluxe.UUCP>
Lines: 13
Nf-ID: #R:aluxe.UUCP:842:uiucuxc:28900016:000:725
Nf-From: uiucuxc.CSO.UIUC.EDU!bantz    Feb  9 09:45:00 1986


[Response to Steven Wong, "AL bell labs" & reply of "friesen@psivax."]

Both the initial request and the response seem to be unaware of the real
purpose of "Eliza."  Joseph Weizebaum (MIT) wrote this program to show that
mere English-like responses which could, individually, be supposed to be
in response to questions or comments of a human, did *not* constitute
natural language understanding let alone "artificial intelligence."  He
took for granted that everyone would recognize that the program did *not*
"understand" anything, or have anything remotely like intelligence.

That eager students of AI took it for more does indeed, I think, show
that the exercise had some real worth, even if only as a cautionary note.