My thoughts

[NumBeast]

Hello,
I just joined this group so please forgive me if I repeat something
that already been said. (this is more of a stream of thoughts than an
actual message)

There are 4 reasons for
1.the evolutionary base of Emotions (if you believe in evolution) is to
short-cut time-wasting rational thought and react immediately to a
potentially dangerous situation. Robots that can acurately simulate
Emotions wil have a beter chance of surviving in a multitude of
situations. This is also an example of f(x) and g(x) both come up with
the same result but one is faster.
2.People are extremely practiced at interacting with other humans.
Robots that can accurately simulate emotions will be better accepted by
the public and can hold jobs with public relations.
3.Emotions (arguably) are the basis of everything that seperates up
from machines. Machines that can feel emotions will be able to perform
jobs currently reserved by humans. Such as Artists or Musicians.
4.for fun
As you can tell at a glance, all these reasons revolve around better
simulation humans. I believe this to be the only reason for simulating
Emotions.

>From a programming stand-point when I think of Emotions the first
things that pop into my mind are Finite State Machines (FSMs) and Fuzzy
Logic. These are the first cooriders I would follow while trying to
simulate Emotions. For anyone who is familiar with these terms it
appears ironic that these 2 ideas are completely exclusive.

Should we simulate Emotions? There is an effect where things that are a
little too human are those that are most feared. If you create an
andriod that is a perfect representation of Humans but have one flaw
your creation will be greatlly feared, It is almost better to stick
with something that is purposely akward. If you create a perfect
replica of Humans is it even a robot anymore? It becomes human. Do you
want the responsibility of recreating and re-defining perfection?

[Ted-san]

Interesting thoughts, thanks for sharing.
Just a note - while this group is infrequently posted to, it is serving
its initial purpose as a repository for interesting ideas on this
topic. High signal to noise ratio is good.

My thoughts on emotion are less that we should be developing robots
with emotion than we should be using what we learn about emotions to
better simulate/model humans or other animals. I agree that giving a
robot animal emotions is where things start getting creepy and
potentially dangerous as well as treading into new moral grounds.

Interestingly, even if the 'simulation' is non-robotic, and just
contained on a CRT, does that make it any less real? It's like the
Chinese box. If we interact with the simulacrum via a computer display,
it could be a program or it could be a real person and if it is in fact
a computer program on the back end that is totally realistic, might we
have the same moral issues of what is life? It all straddles lines
between biology, comp-sci, psychology and philosophy (and probably
numerous other areas).

For years, I've had a fascination with how the combination of simple
rules and chaos creates emergent behavior that can be exceedingly
complex. I see emotions as another emergent facet of the same
phenomenon. Create a complex system and add a layer to modify control
of the other sub-systems and that becomes the emotional component.

Interestingly, while humans have a limited gamut of emotions, it is
certainly possible for the simulacrum to have altogether different
emotional dimensions. I've not thought about this before, but why not?
Just as animal vision only recognizes certain aspects of the world
(i.e. ranges of frequencies in the electromagnetic spectrum) there
exist characteristics outside the range of normal perceptions that are
just as real.

A few thoughts to chew on....

[NumBeast]

Talking to a computer through a console... Ever heard of the turing test? A very interesting read. There is no way of knowing whether a response came from a computer of a human.

 

If you think Chess is cool, wait until you learn about GO! There are about 5 rules to the game of GO. You become an expert by studying your entire life. And computers are no match for even amatures. Talk about complexity. (Sorry, I thought for sure your post said something about chess with simple rules and complex strategy. I decided to keep this paragraph in because it demonstrates a point.)

>Create a complex system and add a layer to modify control
>of the other sub-systems and that becomes the emotional component.

Emotions are not after-thoughts. You do not create a system and then tack on restrictions. You start with an emotional sub-structure and build rational thought onto it.

--
All signatures are false.

[Ted-san]

The Chinese Box is a rejoinder to the Turing test and is an argument
against the strong AI implication that a machine passing the Turing
test is itself intelligent.

Adding a layer does not mean it's an afterthought. The cerebral cortex
is an added layer to the brain that is thought to play a key role in
what makes higher animals intelligent.

The point of my post is that there's essentially no difference between
physical robots and virtual robots as far as ethical implications. The
manifestation of intelligence and emotions brings forth the same sticky
problems you raise regardless of form. "Pulling the plug" on a computer
AI may be no less murder than shooting a perfect robot AI being.