From: Anne Corwin (email@example.com)
Date: Sun Dec 02 2007 - 20:29:59 MST
Stathis Papaioannou <firstname.lastname@example.org> wrote:
Yes, but it's risky applying this analogy to the relationship between
AI and adult human. I don't want to be forced to do things on the
grounds that I am not intelligent enough to know what's good for me,
or even on the grounds that that is what I would wish to do were I a
more intelligent and better informed version of myself.
I am in total agreement on this. Most people are willing to tolerate rather extreme levels of risk in exchange for maximum freedom to self-determine (within the confines of physics, of course -- let's not get mired in arguments about "the illusion of free will" or "the illusion of the individual self"!).
I keep hoping the whole "let's build a Robot Mommy to take care of us!" framing of FAI is just a caricature. Shouldn't advanced AI be put toward important things like oh, say, helping humans deal with the possibility of large asteroids destroying the planet, rather than toward enforcing "bedtime" rules?
I mean, let's keep some perspective here, people. Nobody likes a micromanager, and I doubt the intention of all this high-level AI talk is to come up with something that will tell each of us exactly what to do, every moment of our lives, on a personal level.
My concept of how humans and AIs might relate in the future is one in which the AI operates alongside us, not "above" us like some kind of pseudo-god figure. A truly useful AI would have no need for monkey abstractions like "authority" and "hierarchy". It would also ideally be impersonal the way that science is impersonal -- e.g., we would not owe fealty to it, nor would it force us to do anything in particular, but it would be there as a resource and would effectively prove itself useful through its tangible effects (in the manner of things like evolutionary theory, physics, and the Internet).
Some of the language used on this list really does remind me of theological speculation on the nature of god(s), which I suppose makes sense considering this entire discourse (aside from any actual coding and machine-building) lies in the realm of metaphysics, but it's still a bit bewildering at times to see so many probable atheists engaging in repeated attempts to harmonize incompatible "omni" attributes in a formulation amenable to practical reality.
Be a better pen pal. Text or chat with friends inside Yahoo! Mail. See how.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:01 MDT