Re: SIAI's flawed friendliness analysis

From: Philip Sutton (
Date: Mon Jun 02 2003 - 20:50:41 MDT

Hi Eliezer and Ben,

> Eliezer said: We can discuss governance issues related to AGI
> development. If you want to discuss Iraq, try the Extropians list.

> Ben said: And -- Philip -- a perusal of the Extropians archives will
> show you how contemporary politics tends to be an attractor, and
> dominate an e-mail list far more than anyone could possibly want it to.
> Hence Eliezer's attitude on the subject!!

Points taken. I don't want to discuss Iraq (in an open ended sort of
way) on this list.

But I think one key issue coming out of Samantha's comments (for me
at least) is - how do you regulate in the public good if the regulator is
captured by dubious or criminal elements.

I don't subscribe to the view that that there is such a thing as monolithic
government - you only have to look at the tension between sections of
the US intelligence community on the one hand and the the controlling
group in the Pentagon clustered around Rumsfelt and Wolfowitz over
whether there really was a weapons of mass destruction threat in Iraq.

The same issue applies in relation to police corruption - the one police
force can have straight and bent individuals and straight and bent units
operating under the same umbrella.

I think it's this sort of model of the public sector as a 'mixture of forces -
'both good and bad' that we need to have in mind when devising
regulatory strategies for AGI.

So in practical terms a key question might be: given that governments
usually have a mixture of forces within them, could an effective AGI
regulatory regime be put in place and how could the AGI regulator be
kept honest and effective - despite the existence of corrupt or dubious
forces within government?

We could also ask the question: what would be lost to the task of
getting safe AGI up and running if we fail to harness the positive power
of government - despite the existence of corrupt or dubious forces
within government?

There's bound to be no magic answers to these question but perhaps it
moves us in the direction of dealing with things practically and
realistically - and in the direction of not rejecting options with a sweep
of the massive generalisation.

Furthermore I think that dubious and corrupt forces are less strong -
even relatively weak in at least some countries. If we won't
ackowledge BOTH that there are dubious and corrupt forces in
government and that there are good forces then we can't develop
strategies that will push our countries in the direction of the latter and
away from the former.

I think this bigger picture is the setting in which we have to frame our
governance strategies in relation to AGI.

Cheers, Philip

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:42 MDT