COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\\M1BASL30\M0BDR40\M2BASI30\.
C00014 ENDMK
C⊗;
\\M1BASL30;\M0BDR40;\M2BASI30;\.
\F0\CSTANFORD ARTIFICIAL INTELLIGENCE LABORATORY
\CDEPARTMENT OF COMPUTER SCIENCE
\CSTANFORD UNIVERSITY
\CSTANFORD, CALIFORNIA 94305
\F1



							January 16, 1974



Mr. Daniel McCracken
4 Inningwood Road
Ossining, New York 10562

Dear Dan:

\J	Jerry Feldman  showed  me  the  October  10,  1973  draft  of
\F2Problem-List  of Issues in Computers and Public Policy\F1.  I have
the following comments:

	1. You don't distinguish between  what  you  think  might  be
forbidden  by  law  or  regulation if deemed undesirable and what you
might deplore but would find improper to forbid. This puts  anyone
(like   me)  with  substantial  disagreements  in  a  very  defensive
position.  Maybe they'll forbid home terminal services on the grounds
that they fragment society unless I get my arguments in right away.

	2.  It  seems to be assumed that information utility services
are a natural monopoly.  In  my  opinion,  this  is  not  so  if  the
communication  facilities are a common carrier, as should be the case.
It is very important that there  be  no  monopoly  in  home  terminal
services.   A  home  terminal  should  be able to use any service.  A
paper elaborating this point is enclosed.  This  paper  worries  that
the  current tendency to hang home information services onto cable TV
tends to create a monopoly by the CATV operator  for  whom  providing
information services will be secondary to his TV business.

	3.  It  is not the business of the ACM to decide whether home
terminals are economically justified. It isn't even the  business  of
congressmen's administrative assistants.

	4.  Whether information utilities would speed up or slow down
the processes of government  would  depend  on  how  the  system  was
programmed  and  this  would  depend  on  who  found one or the other
advantageous - the same as now.

	5. The  presently  programmable  abilities  of  computers  to
generate  personalized letters, etc. are a trivial advantage. I would
bet that even  the  best  financed  candidates  spent  only  a  small
fraction of their campaign budgets on the use of computers.

	6. I take a dog-in-the-manger attitude towards an ACM document on the
computer as a metaphor for human self-understanding -  whatever  that
might mean.  Namely, I don't think Weizenbaum has made a contribution
to the matter, but he has a right  to  his  opinion  and  to  publish
articles expressing it.  I would object, however, to his getting his
opinion made ACM doctrine; the ACM has no need for a doctrine on  the
subject.   I  also  have  no  stomach  for  politicking  the ACM into
promoting  my  views  on  the  social  significance   of   artificial
intelligence.

	7. The general public is not as stupid about computers as the
document presumes.  Computers have been around for 20 years and  have
had limited impact on daily life.  The man-in-the-street has concluded
that he doesn't need to know more about computers for the time being.
He  is  right  in this, especially because knowledge of computers has
never been relevant to deciding among candidates for public office--
unlike say a knowledge of economics.  When the situation changes, the
public will show greater interest.
The public would vote for better candidates if it new more about
economics or if it understood the general role of technology better.
I am not clear about what areas of science and technology
most require increased public understanding, but I think it isn't
computers and would not like to compete for attention with other
educational efforts.  On the other hand, there is considerable public
curiousity about the social importance of computers, and it is reasonable
for the ACM to do its part in satisfying this curiousity.

	8. Computer dating is a minor fad.  The computer profession has
no  more responsibility for it than the automobile industry for
the content of bumper stickers.

	9. For poor countries, computer  technology  is  of  less
relative  importance than for advanced countries. Carrying produce to
market on trucks rather than on animal backs has greater payoff,  for
example.

	10. Perhaps the ACM should have a policy on whether CAI should
be supported.  However, this would not be easy to do properly and
fairly.  Thus, suppose an ACM appointed \F1ad hoc\F1 committee decided
that the present CAI projects were losers and furthermore couldn't think
of any ways in which CAI could be profitably be pursued.  Suppose further
that the ACM executive committee reviewed this finding and agreed.
Wouldn't the ACM be obliged to be ready to appoint a committee to review
this opinion whenever a respectable researcher claimed to have thought
of a new good idea in the field.  The NSF is staffed to review proposals
promptly, but even NSF doesn't maintain an official position about
all the fields it hasn't found acceptable proposals in - perhaps
because it is not prepared to review such positions sufficiently
promptly.

	11. The rest of the computer industry doesn't want IBM broken
up at the price of having the  pieces  engage  in  free  competition.
When  the elephants fight, the mice get stomped.  Instead,
whether they know it or not, they are moving towards getting
the government to create a cartel whereby pieces of  the  market
get  parcelled out. This is the significance of IBM's sale of Service
Bureau Corporation to CDC.  The creation of a cartel will not
advance computing in any way.  The main thing that IBM
should be forced to give up is its habit of secrecy.

	12. Whether the Representatives will have to be on the  floor
in the future will be decided by political considerations without the
help of ACM.

	13. The ACM should not go on  record  in  opposition  to  the
First Amendment as would be involved in forbidding the publication of
forecasts.
I can't think of all the effects an attempt to enact such a ban would
have, but we can imagine having to extend it to forbid hints by
candidates that they are in posession of polls that show the public
agrees with them.

	14. I believe the \F2Defense\F1 was not the 701, but I am not
sure.   It  was  the  Mansfield  Amendment (I think) to the Defense
appropriation bill, and I think the year was 1971.

	15. If you think my attitude is  mainly  negative  about  ACM
formulating  policies  in  the public policy area, you are right. The
one area where I would like to see a policy  concerns  antitrust.   I
would  like to see ACM try to formulate the policy that would advance
the technology as rapidly as plπsible.  In this, ACM  should  act  as  a
party  with  an interest in the matter rather than trying to look out
for all interests.  Perhaps there are other such areas where the
computer profession is an interested party.

	16. I think the ACM publications don't have enough on some
aspects of computers and public policy - for example the competition
situation and reviews of computers and major software such as
IBM operating systems.  The situation might be improved by having
the \F2Communications\F1 solicit reviews in these areas.\.


							Sincerely,



							John McCarthy
							Professor of Computer Science
							Director, Artificial Intelligence Laboratory

P.S. You may circulate this if you like.