From: Ben Goertzel (ben@goertzel.org)
Date: Tue Dec 13 2005 - 15:47:10 MST
Regarding an AI without goals: We need to distinguish explicit goals
from implicit goals.
Any nonrandom dynamical system can be said to have an "implicit goal",
so long as its dynamics are such that it maximizes some function.
On the other hand, only some systems explicitly contain internal
models of their own behavior that involve concepts such as "goal."
The "goal" concept in a system's internal self-model does not
necessarily correspond to its implicit goals.
One may ask whether it is possible for a system to achieve a high
degree of intelligence without having an internal self-model that
involves a concept somewhat like "goal". I suspect that the answer is
NO. But I have no proof of this.
However, it's obvious that this internal self-model need not be
preprogrammed; it may evolve as a consequence of the AI system's
interaction with the world and observation of its own behavior. This
is what seems to happen with young human children.
-- Ben Goertzel
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:54 MDT