Re: [sl4] Re: More silly but friendly ideas

From: Charles Hixson (
Date: Thu Jul 03 2008 - 12:49:53 MDT

On Thursday 03 July 2008 12:27:32 am Stathis Papaioannou wrote:
> 2008/7/2 Charles Hixson <>:
> > To say "slave AI" is to select one particular subset of the possible
> > "friendly AIs" and claim "That's the real set!". Those can exist (and I,
> > personally, feel that they are not only immoral, but dangerous to
> > create), but that's hardly the full range. (I'd actually prefer to
> > exclude most "slave AI"s from the category of "friendly", as I don't
> > think that most of them would be. Frustration, at least among mammals,
> > tends to minimize friendly interactions.
> I guess you read "slave" as meaning an AI that is somehow forced to
> serve us even though it would rather not. Do you still think it would
> be immoral if the AI were designed so that it wanted to serve us?

Not immoral, just misuse of the language. (Not a strong misuse, as "slave the
drive to the controler" is standard usage. But slave seems to imply that the
device wasn't changed to make controlling easier.) The problem I have with
that usage is that when the term "slave" is applied to a self-aware entity it
seems to mean that it must obey whether it will or no. The useage you seem
to suggest in "slave AI" being read as if it were a relay or a disk drive
would appear to denigrate the individuality of the AI, which would definitely
be unwise if it had ANY sense of self-preservation (and arguments from many
directions imply that it would...whether built-in or derived). If what you
mean is just what I'm calling "misuse of the language", then the main problem
is how it would shape *your* thoughts, and thus your actions. In that case I
don't consider the usage immoral, only stupid. (Note that these "results"
can't be seen physically in the absence of an AI...but it still shapes how
you think about it.)

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:03 MDT