RE: Military Friendly AI

From: Ben Goertzel (
Date: Thu Jun 27 2002 - 18:15:22 MDT

It's a complex issue, really.

Some aspects of post-traumatic stress syndrome may carry over to AGI's,
others certainly will not...

A big difference though is that, if human programmers notice symptoms of
post-traumatic stress (or other psychological disorders) in an infrahuman
AGI, they should be able to go into its "brain" and take a decent stab at
fixing things. We can't do that with humans yet.

While it's almost certain that AGI's psychological disorders will not
closely resemble those of humans, we really have no way of knowing at this
point the extent to which such disorders will exist. The human brain has
both an irritating propensity to fall into such disorders, AND a lot of
built-in failsafe mechanisms to mitigate against this propensity... an AGI
may lack humans' biological factors that push them toward mental disorders,
but it may also lack sophisticated failsafe mechanisms to help prevent
this.... This will be a really interesting area for research ;)

ben g

> > Which part? Are you saying that humans don't experience detrimental
> > effects later in life related to early childhood emotional trauma, or
> > that this is not applicable to AI. If the later please explain why.
> Did you read section 2 of "Creating Friendly AI", "Beyond
> Anthropomorphism"?
> It's not overly technical, and if you haven't read it, well...
> you really
> should.
> --
> Eliezer S. Yudkowsky
> Research Fellow, Singularity Institute for Artificial Intelligence

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:39 MDT