distributed friendliness built-in OR a friendliness demi-god?

From: Philip Sutton (Philip.Sutton@green-innovations.asn.au)
Date: Fri Jun 04 2004 - 01:22:43 MDT

I'm wondering what's happened to the 'friendliness' project. Early on it
seem that the Singularity Institute was urging AI producers who were
aiming for high and rising general intelligence to build friendliness into
their artificial intelligences. Now it seems like the project has morphed
into the development a single super-intelligence that will divine what
society should do to be friendly and then..............what?

- Publish a paper so we all have access to this wisdom?

- Create a super-powerful (post-singularity?) Power that will impose
    its friendliness formula on everyone and everything else?

If it's the former, that's terrific but it's unlikely to save the world by itself.

If it's the latter then I think the whole thing has degenerated into another
experiment in totalitarianism.

I think the original idea of building friendliness into *each*
advanced/general AI project is the way to go. I figure it's better to seek
a friendly universe as an emergent property of multitudes of friendly
intelligences (biological and non-biological) rather than thinking one
Power is going to impose it.

Cheers, Philip

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:47 MDT