Re: AI Goals

From: Mikko Särelä (msarela@cc.hut.fi)
Date: Wed Apr 26 2006 - 12:26:59 MDT


On Wed, 26 Apr 2006, Jef Allbright wrote:
> Woody, I am saying that it is necessary that we increase our awareness
> of *both* aspects: our evolving subjective values and our evolving
> scientific/technical knowledge. In order to make "good" decisions, it
> is necessary to understand which direction we want to go, and what works
> to get us there.

Hi,

I assume you are familiar with game theory from the way you write. Have
you read the things that exist in the intersection of morality and game
theory, such as David Gauthier? The following link might provide some
useful stuff for you.

http://plato.stanford.edu/entries/game-ethics/
 
Some other things that might be of interest include Karl Popper and then
David Deutsch. If you haven't read the latter's Fabric of Reality, you
might find it interesting. It probably doesn't contain a lot of new stuff
for you, but it does put things together in a fresh and interesting way.

I've been working through a lot of game theory and morality stuff in my
past and have found that an approach that is very good for understanding
things.

As you might know, there are people who believe that moral systems
cannot be distinguished, because value propositions are claimed to be such
that no natural thing that happens can justify, or refute them. This
actually does not hold, if we require a moral system to describe both
values (goals) and the means to get there.

One of the simple rules that David once suggested for determining goodness
of such a moral system is to determine whether it succeeds by its own
lights, i.e. do the means accomplish the ends that the moral system
advocates. Note that this is a necessary, but not sufficient condition for
a good moral system. We should be able to develop other good ways of
critisizing moral systems and thus evolve our knowledge.

-- 
Mikko Särelä	http://thoughtsfromid.blogspot.com/
    "Happiness is not a destination, but a way of travelling." Aristotle 


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT