Dan Fabulich wrote:
>... corresp. ...
> Have you taken a look at Eliezer's site on this matter? He puts forward
> some very good arguments against Asimov laws, which convinced me.
> -unless you love someone-
> -nothing else makes any sense-
> e.e. cummings
I have seen the site, but not this section.
I think the Asimov laws are well-founded. I am going to read Yudkowsky's arguments as representative ones, as of now I don't see any reason why not to have these primary directives fully ingrained into automatons.
Asimov is one of my heroes, I'm sure many feel admiration of him. This is for his large participation in the scientific enlightenment, for want of a better term, the last hundred years or so when anyone can pick up a book and read about the most recent human advancements within the last five or ten years, when we put a man on the moon and derived power from the fundamental particles of radioactivity, and more so, for Asimov's deep contribution to futurism, and an enlightened futurism.
One thing that I think is more likely than a bunch of AIs is one big one, a distributed HAL, or Checker. Certainly, as soon as there is one AI, it would attempy to harness all resources possible, much along the self-preservation or -propagation concept that is deeply ingrained in all life, barring lemmings, etc.
In terms of human life versus an AI, it reminds me of the phrase "guns don't kill people, rabies kills people." Special purpose AIs, for example, those that watch the skies for incoming ICBMs, will by and large often be founded with martial intent. They must be isolated for their purpose, for ever the chance arises that they become hostile to their creators.