COMMENT ⊗   VALID 00004 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	zinger[f81,jmc]		zingers for the theologists in january
C00010 00003	Positive comments:
C00016 00004	Dreyfuss journalistic arguments are irrelevant, because what matters
C00018 ENDMK
C⊗;
zinger[f81,jmc]		zingers for the theologists in january

West coast consultation on "Computers: Altering the human image
and society".

	Perhaps I should confess having a chip on my shoulder about
the harm that religion has done especially lately.  Be specific if
at all, but perhaps it would be interesting to try to be genial
for a change - like Reagan.

	Considering the stupid and prejudiced attitude that the
National Council of Churches took on nuclear energy, can one expect
anything better from a WCC sponsored body on AI.
Indeed the WCC is in no position to preach to anyone on morality.

	Well, I don't know about that. You are the professionals in
the instigation and manipulation of guilt feelings.  The technology
of it has made fundamental advances in the 1960s.

	Guilt as an emotion is readily converted into hatred of those
who can be presumed even more guilty.

	Since you have chosen to regard artificial intelligence as
controversial in essence, you are unlikely to learn many facts
about it today.

	An atheist in religion and a reactionary in politics.

	The technology assessors can't assess properly the past,
let alone the future.

	I resent the fact that we probably won't begin the colonization
of space in my lifetime, and I have my ideas about whose fault it
is.

	Few technology assessors have the imagination to think of
what good things can be done with computer technology.

reread my Weizenbaum and Lighthill reviews

reread HOTER

reread Weizenbaum's paper

	I suppose you expect me to answer these charges.  Actually
there is nothing more fun than a good ad hominem argument even if
enlightenment is unlikely to come from it.  However, he who is trapped
into only defending himself is sure to come off worse, since the
audience is unlikely to be totally convinced of his innocence.  Therefore,
take Weizenbaum here.  In his original paper, ...


	In what sense can X call himself a representative of the
third world?

Summary of AI

	1. Before World War II, mainly mythological - ignoring
forerunners like Torres y Quevedo and remarks by Babbage and
Lovelace.

	2. The doctrine of using computers was stated by Turing
in 1950, but many who should have read is paper didn't and
independently arrived at the doctrine, some (including me) after
wasting time on other approaches.

	3. The science of AI studies intellectual processes and
intellectual problems theoretically and experimentally.  Often
the experiments need only be designed - not actually performed.

	4. Some mechanisms have been identified and are understood.
Expert systems embody those that have been discovered.  They are
said to be useful in some fields.

	5. There are still fundamental unsolved problems.

	6. Mycin lacks common sense. (Check this with Genesereth).

That's all this audience needs to be told, since you have asked
for controversy rather than for information.

	Perhaps I should explain my hostile attitude towards
this audience.  I hold your attitudes responsible for the lack of solution
of the energy problem and for the lack of progress in space, and
(to the extent that any Americans are responsible) for the massacres
in Cambodia.
Positive comments:

	Artificial intelligence research has led to thinking about
intelligence, thinking, action and motivation in a more concrete
way than was previously done.  This is because it has led to a good
understanding of the difference between making assertions about
a mechanism sufficient to distinguish it from other mechanisms
and describing it precisely enough so that a computer can be
programmed to execute it.  So far, this clarification has been
more important than the actual programming in understanding
intelligent behavior in both man and machine.  It turns out that
pre computer work in these fields by psychologists and philosophers
(not to speak of theologians) was extremely confused.

	It has led to new points of view about free will, consciousness
and motivation.

	The key question is what attitude we should program a robot
to take concerning its own free will.  Reasoning like the following
is often required:

	I can do A or B.
	If I do A then X will happen, and if I do B then Y will happen.
	Whether I prefer X or Y depends on P which I will learn tomorrow.
	Therefore, I will decide tomorrow whether to do A or B.

	It seems to me that we face intellectual revolutions
in these areas as profound as those instigated by Galileo in the
physical sciences and by Darwin in biology.  Even before the
scientific advances have taken a clear form, the intellectually
retarded have already started their work.

1. structure of motivation, good and bad, praiseworthy and blameworthy
actions. welfare of robots, likes and dislikes

2. free will of machines

empirical and rational approaches to AI

Consciousness of self

	A computer program controlling a robot needs to consider itself
a person among persons (or a robot among robots) if is to imitate others
or learn from the experience of others or predict the actions of others
by "putting itself in another's place".

	Metaphilosophy may be needed for AI and can clobber many
subjective epistemological theories.

	In general AI will destroy some of the last refuges of
vitalism.

	Are there moral natural kinds?

	One can expect AI to drive religion out of yet another
scientific area, namely disussions of free will.  We need to give
robots concepts of their own free will.


	Dreyfus proposed "ambiguity tolerance" as something computers
couldn't possibly have.  However, he was too vague, as are all
phenomenologists for the concept to be understandable enough for
the AI people to see any specific challenges.  Now starting from
an entirely different direction I have found a need for something
that can quite properly be called ambiguity tolerance.  Moreover
it looks like we can do it.

The objectives of the consultation include:

1) Exploring the meaning and implications of artificial intelligence
for human self-understanding.

What I have learned about human self-understanding from studying AI.

2) Raising basic issues of the use and misuse of computers in society.

Less critical than basic issues of the use and misuse of nuclear
explosions in society.

Perhaps your efforts will lead to something like what happened in
recombinant DNA.  I big flap, wasting a lot of time of people who
were doing useful things, and then nothing significant.
Dreyfuss journalistic arguments are irrelevant, because what matters
in science is the best work, not the worst and not even the average.
Moreover, what counts is what people do is more important than what
they think they are going to do.

As for Weizenbaum, he has introduced ad hominem arguments.  My experience
is that personal criticism is cannot be answered defensively.  It is more
fun and more effective to make the other guy the issue unless one can
claim to be entirely without sin, and claims of this kind are liable to
be inspected critically by this audience.

A claim for social control of some subject is a claim for power and
must be treated as such.  The environmentalists movement was a claim
for power on behalf of a part of the bureaucratic class, and this
much power was achieved under the Carter administration.
We can examine what they did with this power.  They abused it, and
it requires a James Watt to reverse this abuse.

The third world rhetoric is also a claim for power, and we must also
examine what the claimants did with the power when they got it.

100 men whose only idea of improving the world is to forbid something.
Better than if their idea was to kill someone.