Re: Mind Reading - Approach to Friendliness

From: Robin Lee Powell (rlpowell@digitalkingdom.org)
Date: Tue Jun 20 2006 - 11:32:43 MDT


On Mon, Jun 19, 2006 at 10:24:54PM -0400, Ben Goertzel wrote:
> >> Concept is pretty simple. You develop really powerful
> >> analytical/debugging applications that can display the entire
> >> contents of the AI mind in tractable and useful visual
> >> representations and extrapolations into future states of the
> >> mind.
> >>
> >> Strategy: Let the AI 'sleep' (i.e. down-time) long enough to
> >> periodically analyze the entire contents of the mind.
>
> This basically means the AI cannot be alllowed to be strongly
> superhuman, because if it were, we humans would certainly not be
> able to understand its thought processes, let alone extrapolate
> them...

Depends on whether there is Turing-completeness for intelligence,
and whether humans have reached it. It may simply take a Really
Long Time for us to understand/extrapolate the AI.

Now, whether each real-time second of AI thinking should be hampered
by a few months of humans arguing about it is a whole different
issue. :-)

-Robin

-- 
http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/
Reason #237 To Learn Lojban: "Homonyms: Their Grate!"
Proud Supporter of the Singularity Institute - http://intelligence.org/


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