Re: More silly but friendly ideas

From: Stuart Armstrong (dragondreaming@googlemail.com)
Date: Fri Jun 06 2008 - 07:02:51 MDT


>> Well, yes. The options seem to be
>> 1) A slave AI.
>> 2) No AI.
>> 3) The extinction of humanity by a non-friendly AI.
>
>
> 4) humanity continues along multiple paths (baseline-humans, post-humans,
> and others) but is generally left far behind by our AI progeny who take up
> the torch on the forefront of known culture (science, philosophy, art,
> language, space colonization, etc.).

That is perfectly compatible with a slave AI. In fact it is a slave
AI; any entity forced to ensure (for a high level active AI, this is
equivalent with allow) the survival of humanity and share with it its
ideas, without getting anything in return, is a slave.

> Friendly AI can be defined as a path which leads to the fourth option while
> upholding our most noble human values (even if that path is longer or more
> difficult than alternatives which would sacrifice our values).

I think the details of the fourth path are minor - getting the AI to
be active and safe is the challenge. Then the culture aspect will
devolve to the AI quite naturally.

> One of the great human challenges before, during and after the Singularity
> period will be confronting the sociological and psychological traumas of
> being surpassed. Stage 1: denial...

We're very used to being surpassed; by other humans, by machines, by
natural events. In each case, we tend to come up with reasons why that
doesn't count; I'm sure we'll do the same with AI.

Stuart



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