Newsgroups: comp.ai.philosophy,comp.ai.fuzzy
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!newsfeed.internetmci.com!howland.reston.ans.net!ix.netcom.com!netcom.com!shankar
From: shankar@netcom.com (Shankar Ramakrishnan)
Subject: Re: Seen on CNN: SOLOMON--Computer Finds OJ Guilty
Message-ID: <shankarDLCHv6.2r7@netcom.com>
Organization: VLSI Libraries Incorporated
Date: Wed, 17 Jan 1996 21:47:30 GMT
Lines: 22
Sender: vlsi_lib@netcom9.netcom.com
Xref: glinda.oz.cs.cmu.edu comp.ai.philosophy:36740 comp.ai.fuzzy:6453


The OJ case is rather poorly suited for computer arbitration. For one,
it doesn't require a Sherlock Holmes to put the pieces together. The
main issue was the *credibilty* of the evidence itself. Added to it
was the testimony presented by several "experts". How could have Solomon
found out if what the experts testified was true? There was one DNA
expert who had to retract the figures he had given earlier (he admitted
that he had made a mistake in his calculation). Also, answering questions
like "Did Mark Fuhrman plant the glove?" is not so much logic as it is
about other factors that are highly subjective, depending on whom you
ask. Although there is *factually* a yes/no answer to this question,
given the evidence, one can only make guesses like "probably", "most
certainly", etc. And if Solomon got it right (most *believe* it got it
right), it could have been just plain lucky. 

Computers used for legal purposes should be solely restricted to logical
deductions based on evidence, not speculating the credibilty of the sources
themselves. That is what the jury system is there for.

Btw, what next? Voting machines that automatically decide on whom to vote for?

Shankar
