From: Nick Bostrom (firstname.lastname@example.org)
Date: Mon Mar 05 2001 - 17:47:25 MST
As I see it the following points have been made in reply to my original
1. Currently the military probably regard SI as too far out to merit
2. As SI draws closer, the military will get about it independently of our
I agree with both these points but they don't really answer whether we
should try to do our little bit in promoting awareness among the military.
3. Eliezer claimed that SI has no military applications
I think this is too categorical. Maybe the designers' experience with a
gradually growing AI will persuade them to adopt some reasonable ethical
standards, which may then converge into Friendliness as the SI matures. But
I can't be completely confident that this will happen. Moreover, from the
point of somebody with non-reasonable ethical views, the prospect of
"friendly" SI could be seen as a military threat. "What will happen to our
religious revolution if those guys develop what they call a friendly (but
probably ungodly) SI?"
4. Eliezer hopes that there will not be an SI arms race.
Ok, so presumably the more awareness there is of SI the greater the
likelihood of arms race. But
5. Eliezer also hopes that if there is an arms race then "protect America"
will be a subgoal of "serve humanity"
And we might think that THIS is more likely given that there has been long
awareness and discussion.
So what's been said doesn't really answer the original question. Now my
intuition that it is better if the military knows results from the
heuristic: When in doubt, promote public discussion.
BTW, since this is my debut at SL4 I ought perhaps to introduce myself.
Those who don't know me from other forums are welcome to check out my
homepage at http://www.nickbostrom.com.
Department of Philosophy
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:36 MDT