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.

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?

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....

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.)

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.

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.

Hi

Thanks for sharing.

This is a fun topic to play around with.

Reading over the past few postings, got some input/thoughts... :)

First of all about the moral dilemma surrounding the question "is artificial
intelligence real life?" and where the limits are.

I'd say - yes absolutely. Limits are at the same place as with biological
intelligence. The problem is that today most of us view it in black and
white, and the rules we as a society set (in norms or by law) is not
compatible with the notion of true AI. We will have to re-consider the value
and limits of real-life intelligence first. Then figure out where to set the
limit for artificial intelligence. To most people this is an difficult task
to accept, often because it is governed by religion and a notion of what
being a kind (self-rewarding) being means.

I'm guessing we will not find and agree upon a limit without having the
technology around for some time first. I think most people will be incapable
of empathy for robots with/or true AI until it becomes more common and some
generations has passed.

So to AI...

I believe that thinking of true AI and robots as the same is necessary.

Simply because a robot capable of being any value and accepted as a "sane"
intelligence would have to be raised in this world similar to a child. Care
and supervision of parents, interaction with the world, making mistakes and
being punished for them, going to school, etc are all key elements. You
cannot create an artificial intelligence program and expect it to understand
anything useful, or make itself acquainted as an intelligence, unless it has
been raised like this. True AI will be the result of amount and type of
input data it has, in addition to the correct algorithms.

There is a general notion that this planet consists of a wide variety of
people. That is true, if you take into account only the human mind and human
upbringing.

The majority of alternative "intelligences" are probably useless to us at
this point in time, and most would probably not be recognized as
intelligence. So to create true AI one has to get it pretty close to home.
Living things have a distinct behavioral pattern because of elements like
their 3D environment, their needs and their genetic programming/physical
existence.

The physical world and our society, with its norms and behavioral patterns,
can be used to reveal some important answers in our quest for the holy
grail. This because it's both a pattern that has been evolutionary adjusted
to fit our intelligence, but also because our intelligence is based on an
upbringing within in it.

3D worlds like Second Life is in my opinion our current best bet for someone
to develop something that resembles true AI. SL allow custom-build clients
to connect (sl-lib), so developers can create their own AI bots. It has a 3D
world with physical limitations and real humans to interact with. As
development of this kind of games gets them closer to reality it becomes
possible to do some interesting research without the need for robotics
skills and a large bank account. The gaming industry is already pushing
development of AI game characters.

Eventually some 3D world will probably get a realistic ecosystem, which
would make AI programming much easier.

Then finally emotions...

Emotions seems to be what largely governs the behavior of both humans and
animals. Decisions are very flavored by emotions. In real-life intelligence
it may seem as if decisions made by the brain are weighted by emotions. We
tend to make unconscious and conscious decisions to experience positive
things and avoid bad things.

But emotions can make people knowingly and willingly make the wrong
decision, even though it will lead to bad things. The simplest example of
this is anger.

Take for example someone who would throw their mobile phone into the wall
breaking it in anger. The immediate assessment that any person is capable of
is that the mobile phone will not work, which will cause a bad feeling and
expense. Still the decision is made, even though this person normally do not
throw mobile phones into the wall. This action can be repeated with another
object, for example the TV-remote or iPod.

So I guess one can rule out that the consequences has not been considered
before the decision is made.

In the same way we see the reverse pattern. Someone receives a mail and
becomes angry. This involves mail being read (optical processing), linked up
to relevant memories and processed, and the emotion being triggered. What
emotion is being triggered is also based on what emotions one currently has.
Possibly as a result on decisions being made during reading.

Looking at it this way it may seem that emotions are an ambient existence
that forms a common ground for more concrete decisions. That one cannot
exist without the other. That emotions are essential in what we consider
intelligence.

Long posting ... :)

Sincerely,

 Tedd

Hey Tedd, Interesting points.
First i would like to say nice vision on the whiole 3D SL theory. I'm
not the greatest fan of it but interesting! I saw a video (from the
BBC) not to long ago that talked about time travel and basically they
talked about how the future can evolve with such worlds and we can
recreate the past and future. Also how we'll never know that we
technically did already, which would mean currently "this" world we're
living in be one of them. Interesting enough but a bit off topic now.
Taking your view of this and their's going the same direction of
predictions is catchy. I do see a great height of technology growth
from this area! and possibly the first real visual interactive AI
emotions to come from it and later transformed into actual hardware as
technology innovates itself. Always worth the conversations.

Overall, good read.
Would like to hear more from your views in the future!

Also, about the limitations of how people see AI. I had a debate with a
friend recently about this. very true! people don't see it and think
very "single minded" as i'll rather say than "out minded" coming out
the box and seeing the future! But once people start to see such
machines produce massive amounts of emotions just like us i'll say
that's when things will start to shake up a bit. Which pretty much
concurs with your statement above.
Another thing that pops in mind when writing this is a BBC article that
was online not too long ago talking about equal rights from AI.
Interesting how european nations are already doing research on getting
such a thing on the ball already. Just a matter of time till sociality
starts accepting things.
Scary part is... so much is still to be discovered.

On Jan 3, 7:34 am, "Tedd Hansen" <tedd.han...@gmail.com> wrote: