Re: The AI revolution

Anders Sandberg (asa@nada.kth.se)
02 Jul 1998 09:30:19 +0200

Hara Ra <harara@shamanics.com> writes:

> Asimov invented the Three Laws as a device by which conflicts could be
> created for his robot stories. A case of simple statements leading to
> impossibly complex results.

I think this was the greatest success of his robot stories: he managed to show how three apparently simple and logical rules could produce very complex and unexpected results. In many ways Asimov was before his time in this respect.

> So, what if a robot has this choice:
>
> Kill someone, and allow 100 others to live, or
> not kill, and allow the 100 others to die.
>
> This would probably immobilize the robot, which is the worst choice,
> so the Zero'th Law is:
>
> 0. A robot, when faced with a choice which results in harm,
> chooses the one resulting in the least harm.

OK, I see the reasoning now. Of course, it took some pretty smart robots to come up with this, as far as I can remember it was just Griscard (?) and Daneel who suceeded.

> Implied is that this Law overrides the other three; that the phrase
> "except when in conflict with the Zero'th Law" is added to the first
> 3 Laws.

This suggests a loophole in the original laws, that you can add higher level laws derived from the others, which was probably not intended by the human designers. Another unexpected feature; I can really understand why Susan Calvin decided to become a robot psychologist.

-- 
-----------------------------------------------------------------------
Anders Sandberg                                      Towards Ascension!
asa@nada.kth.se                            http://www.nada.kth.se/~asa/
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y