Theoretical question about AI

Post Reply New Topic

I understand that scientists currently do not understand everything about how the human brain works. But say in the future if or when scientists understand everything about how the human brain works, and are able to artificially construct one and place it in a robot, would that mean that robot would be able to have human emotions and feel things such as pain?

I apologize if this is a stupid question, I am just curious.

From here, it seems entirely likely. Of course, such an AI might well be constructed out of biological cells...

Probably not. Emotions are chemical reactions.

^^ But do not these chemical reactions and messengers trigger electronic feedback and responses in the nervous system and brain? and are these responses not already being manipulated by electrodes as in during brain surgery and treatment of brain disorders? so could not the chemical part of the cycle eventually be replaced by electronics, bioelectronic or otherwise.

I recently read about the invention of a fully biological (bacteria + virus) electrical cell (except for the contacts, of course).

Anything that the human mind can conceive is possible ... except for striking a match on a marshmallow, or putting toothpaste back in its tube.

you need to distuniguish the difference between a copy and a artifical version
an exact copy of lets say my brain, means taht they will both respond in the same way to the same stimuli

given ofc that the events we/they perceive are in paralal too

an experemnt was done in 2006 with MIT -

-rat brains + flight simulator

it laenred that based on cirain imputs it could eventually fly the plain,

it learned which ones ment crashes (that hurt it) so leanred to avoid things

it leanred how to control the plain basically.

Ive gtg now sorry for how breif this post is

Yes.

You can always simulate emotions and things that make a character happy an sad. Just look at "The Sims" characters.

As for sensations (pain, touch) you could do that with touch/motion/other sensors that gave feedback to the "brain".

I think that androids will get so complex and sofisticated that you will never be able to distinguish who is what unless you opened one up - but frankly; i would not care if my best friend/coworker/soulmate was an android, as long as everyone was happy.

Maby there could be some sort of Voigt-Kampff test, that could see through the android responses and could tell who is who:

http://en.wikipedia.org/wiki/Voight-Kampff_machine

But even with such a machine, it would be only a matter of time before engineers would make an even better android that could pass that test.

Athough at the end of the day
what is pain

Negative signals that warn us about stuff.

So pain = something bad

AI are already able to follow this concept and improve to "avoid" "pain"

Alec

What if the robot used electricity to travel from its artificial brain to the rest of it instead of neurons? Assuming the robot also had some sort of system that acts similar to the human nervous system.

I think the real changeling part is the concious;

We still don't understand why we have this self awareness, and to which extent this self awareness is present in animals (our next relatives seem to have at lleast a kind of self awareness). To simulate this self awareness, which is beyond the more "mechanistic" function of emotions, we perhaps need to understand first how the brain does produce this self awareness to simulate it.

---

That emotions are not really attached to self awareness we aspies are the best example - we are self aware, but our emotional system works differently than with NTs.

It is certainly possible for people to feel empathy for a machine. Hopefully machines can feel it back.

Check this site out. It put a smile on my face.

http://www.tweenbots.com/

I see no reason why a sufficiently advanced AI should not be self aware and conscious and experience emotions, pain and pleasure. Of course there will be those who say they are only mimicking emotions and not really "feeling" them, but I think that "feeling" is likely synonymous with processing certain types of data anyway.

It may seem an outrageous thing to say but I would not be surprised if the microprocessors in today's computers feel "something" - not in the human sense of the word, but more on an abstract level.

I think "feeling" is simply an emergent property - an abstraction of high level data recognition and manipulation.

It may be that advanced AI designs will destroy the illusion of self awareness and free will in humanity.

Here's an article from BBC news about a robot with a rat brain. While it appears very far from experiencing emotion, although how would we really know that, but in a simlar vein to the flight simulator quoted earlier, it learns to avoid walls and such. Very interesting!

A rat brain is several orders of magnitude in advance of any artificial robot.

ruveyn

I still say true AI can only be accomplished with a learning machine. You can't be created and know everything straight away. If that were possible, think where we might be today. We have to learn everything.

Pain is just a way of telling you that part of the body is damaged and must not be used so healing can take place

Excellent point. We have to approach AI from the standpoint of a child if we are going to crate something highly intelligent within this lifetime. Yes, we may be able to programme it to do anything, but to LEARN anything, that is the difference. And no matter what, that is the pinnacle that should be aimed for, a robot that can survive independent of it's creators, intelligent enough to learn. Though, then you have Skynet.